**4.2 Human artificial chromosomes**

Human Artificial Chromosomes (HACs) are vectors able to replicate and segregate in parallel with the endogenous chromosomes in human cells. To achieve this, they must contain the minimal elements required for chromosome function, namely an origin of replication, telomeres and centromeres (Pérez-Luz & Díaz-Nido, 2010).

One approach towards constructing HACs, called "top-down", involves fragmentation of already existing chromosomes and generation of smaller mini-chromosomes, where only the three functional chromosomal elements remain. Several studies have shown that minichromosomes can host and allow the expression of large therapeutic genes, be transferred between various mouse and human cell lines and be transmitted through the mouse germ line (Kakeda et al., 2005; Shen et al., 2001; Voet et al., 2001). Though mini-chromosomes have useful properties for application in transgenics, their use in gene therapy is restricted to an *ex vivo* approach only.

HACs with a high potential for use in *in vivo* gene therapy are generated by a different approach named the bottom-up. This is similar to the method applied for YAC construction in yeast and involves assembling the functional chromosomal elements and building up a HAC *de novo* in human cells. Different strategies have been followed to generate *de novo* HACs, the most convenient of which is to transfect a BAC carrying only a large array of αsatellite (alphoid) DNA and some marker genes into HT1080 cells. No telomeric sequences or an origin of replication have been shown to be required, probably due to generation of circular HACs and initiation of replication at origins found within the marker genes (Ebersole et al., 2000). HACs generated this way exist as single (or low copy) chromosomes in the nucleus and have a high mitotic stability (close to 100%) in the absence of selection. The potential use of these vectors in gene therapy has been demonstrated by expression of large therapeutic genes from them (Grimes et al., 2001; Mejía & Larin, 2000). Further advance was noted when efficient methods for manipulating large sequences of repetitive nature, such as alphoid DNA, were developed and used to generate HACs, as will be discussed later on (Kotzamanis et al., 2005). Nevertheless, several issues need to be solved before any clinical application. First, HACs have been shown to form efficiently only in HT1080 cells so far. Whether this is due to their inability to form in other cell lines has not been answered yet, but is limiting their use for *in vivo* gene therapy. Second, all HACs produced by *de novo* synthesis in HT1080 cells have been between 1 and 10 Mb in size, definitely smaller than native chromosomes, but larger than the input DNA, suggesting that unpredictable amplifications and rearrangements have occurred during their formation, which is not desired for safe gene therapy vectors. Third, other fates of the input DNA than formation of HACs have been observed and integration has not been excluded (Harrington et al., 1997). This would not be a problem in *ex vivo* therapy where individual clones expressing the gene of interest from a HAC could be isolated, but in *in vivo* gene therapy, any interaction of the input DNA with the endogenous chromosomes could have the same consequences as viral vectors have. Further research is necessary to increase the efficiency of HAC formation so as to ensure that no integration events take place.

#### **4.3** *S/MAR* **vectors**

10 Non-Viral Gene Therapy

The best characterized system for episomal maintenance is based on sequences derived from the EBV genome. EBV is a member of the herpesvirus family with a 172-kb genome that is latently maintained as an independently replicating episome in a small percentage of infected lymphocytes (Masucci & Ernberg, 1994). During the latent phase of its cycle, DNA replication occurs from the origin of replication *oriP* and only about 10 proteins are produced of which the only protein that is required for replication at *oriP* is the Epstein Barr Nuclear Antigen-1 (EBNA-1) (Yates et al., 1985). The interaction of *oriP* with EBNA-1 also enables the segregation of the viral genome between the daughter cells through the

These features of the EBV have been exploited to develop a system for episomal maintenance of foreign DNA delivered into cells. It has been shown that plasmids carrying *oriP* and expressing EBNA-1 can replicate autonomously once per cell cycle when delivered into human cells and can segregate by attaching to the host chromosomes (Haase & Calos, 1991). The *oriP*/EBNA-1 system has also been shown to support long-term episomal maintenance without selection and expression of very large human genes, such as the *CFTR* (Huertas et al., 2000), the human hypoxanthine phosphoribosyltransferase (*HPRT*) (Wade-

results, a convenient system for adding the *oriP*/EBNA-1 sequences onto any BAC already containing a therapeutic gene has been developed (Magin-Lachmann et al., 2003) and will be

The *oriP*/EBNA-1 retention system is easy to use and can provide extra-chromosomal maintenance to foreign DNA of hundred kilobases delivered into cells but has some major disadvantages. It provides random rather than equal segregation of the episomal vector to daughter cells which results in loss of the episomes at a rate of 2-8% per cell division (Sclimenti & Calos, 1998). This, along with the fact that it involves viral sequences particularly from the EBV which has been associated to several types of human

Human Artificial Chromosomes (HACs) are vectors able to replicate and segregate in parallel with the endogenous chromosomes in human cells. To achieve this, they must contain the minimal elements required for chromosome function, namely an origin of

One approach towards constructing HACs, called "top-down", involves fragmentation of already existing chromosomes and generation of smaller mini-chromosomes, where only the three functional chromosomal elements remain. Several studies have shown that minichromosomes can host and allow the expression of large therapeutic genes, be transferred between various mouse and human cell lines and be transmitted through the mouse germ line (Kakeda et al., 2005; Shen et al., 2001; Voet et al., 2001). Though mini-chromosomes have useful properties for application in transgenics, their use in gene therapy is restricted to an

HACs with a high potential for use in *in vivo* gene therapy are generated by a different approach named the bottom-up. This is similar to the method applied for YAC construction in yeast and involves assembling the functional chromosomal elements and building up a HAC *de novo* in human cells. Different strategies have been followed to generate *de novo* HACs, the most convenient of which is to transfect a BAC carrying only a large array of α-

malignancies (Cohen, 2000) limits the use of EBV vectors for safe gene therapy.

replication, telomeres and centromeres (Pérez-Luz & Díaz-Nido, 2010).

*-globin* gene (Black & Vos, 2002). Following these promising

association of EBNA-1 with host metaphase chromosomes (Harris et al., 1985).

β

**4.1 OriP/EBNA-1 episomal vectors** 

Martins et al., 2000) and the

**4.2 Human artificial chromosomes** 

analyzed later on.

*ex vivo* approach only.

*S/MARs* are diverse sequences found in all eukaryotic genomes where they are involved in many aspects of chromatin function such as organization of chromatin into loops, which seems to be mediated by the interaction between *S/MARs* and the nuclear matrix (Heng et al., 2004). Experiments with a plasmid vector containing an *S/MAR* element isolated from the human interferon β*-* gene has revealed one more feature of *S/MARs*, their ability to provide episomal maintenance of foreign DNA introduced into cells. This vector was able to replicate and remain episomally in CHO cells at low copy number for more than 100 generations in the absence of selection and with a mitotic stability of 98% (Piechaczek et al., 1999). It was later confirmed that the mitotic stability of the vector was provided through the interaction of the *S/MAR* with the nuclear matrix (Baiker et al., 2000). Interestingly, the *S/MAR* used seemed to prevent vectors from integrating into the host genome as integration events were observed in less than 1% of stably transfected clones (Jackson et al., 2006). The ability of the interferon β *S/MAR* to provide replication and episomal retention has been demonstrated in several cell lines and in primary cells (Papapetrou et al., 2006) and also *in vivo* in genetically modified pigs (Manzini et al., 2006). Furthermore, the same *S/MAR* element has been introduced by site-specific homologous recombination to a BAC carrying 135 kb of the human low density lipoprotein receptor *(LDLR)* genomic locus and shown to provide low copy episomal maintenance in CHO *ldlr*-/- cells for more than 100 generations without selection and long-term expression of the transgene at high enough levels to completely restore LDLR function in these cells (Lufino et al., 2007).

In summary, it seems that unlike viral episomal vectors which need to encode viral factors required for their function, S/MAR vectors achieve their replication and segregation by

Non-Viral Gene Therapy Vectors Carrying Genomic Constructs 13

Some features of the Adeno-associated virus (AAV) can be exploited for site-specific integration of foreign DNA. AAV is a non-pathogenic virus with a 4.7 kb single-stranded DNA genome comprising two genes, *rep* and *cap* flanked by 145 bp palindromic sequences termed inverted terminal repeats (ITRs) (Srivastava et al., 1983). In the presence of a helper virus such as adenovirus or herpes simplex virus, AAV can undergo replication and enter its lytic cycle while in their absence AAV integrates into the human genome and becomes latent. The ITRs contain the sequences required for replication, packaging and integration of the virus and the *rep* gene encodes four regulatory proteins required for catalysis of integration into the human genome during latency. Integration occurs into a specific site on

It has been shown that human transgenes flanked by ITRs can integrate into the AAVS1, with the minimal requirement for expression of viral Rep in cultured human cells (Cortés et al., 2008), and *in vivo* in transgenic mice carrying the human AAVS1 (Liu et al., 2010; Recchia et al., 2004) . The AAV system has also been shown to be able to integrate large genes of 100kb in size into the AAVS1 (Oehmig et al., 2007). These studies have demonstrated the effectiveness of AAV-based vectors but again safety issues have emerged. Integration of small transgenes has been detected in 10-30% of infected human cells in culture with only about half of the integrations occurring specifically at the AAVS1 (Recchia & Mavilio, 2006), suggesting that there would be a 50% probability of insertional mutagenesis in a gene therapy application. In addition, persistent expression of the viral Rep protein is toxic and can cause chromosome instability and mobilisation of the transgene

Gene therapy using genomic constructs entails engineering of large DNA fragments often of repetitive nature. For instance, marker genes and other useful sequences, able to confer extra-chromosomal maintenance, need to be added to vectors carrying large genomic fragments, without causing any rearrangements. In other cases, the entire genomic locus of a therapeutic gene may not be available in a single BAC vector and linking of the inserts of two or more BACs is necessary. The technical difficulty in performing such manipulations has hampered progress in this field for a long time. Only recently, efficient engineering methods have been developed allowing the consideration of BACs carrying genomic loci as

In most cases, the cloning of mammalian selectable markers and small reporter genes on the vector region of a BAC carrying a genomic insert by classic molecular biology procedures is limited by lack of convenient cloning sites on the vector, the possible presence of many restriction sites in the insert and the difficulty in manipulating large DNA molecules without affecting their integrity. Alternatives to classic molecular biology techniques using restriction enzymes are therefore required. For adding reporter genes and short sequences onto BACs, site specific homologous recombination mediated by the bacteriophage P1-

The Cre protein recognizes and catalyses efficiently recombination only between specific *loxP* sites, which are present on all BAC vectors making a modification method based on them

**5.2 Adeno-associated virus based vectors** 

chromosome 19 called the AAVS1 (Kotin et al., 1992).

**6. Methods for modification of large DNA molecules** 

(McCarty et al., 2004).

gene therapy vectors.

**6.1 Addition of marker and small genes** 

derived Cre/*loxP* system is such an alternative.

recruiting and interacting with host cell proteins. Moreover, unlike HAC vectors, *S/MAR* vectors do not need to undergo amplifications and rearrangements to function and are therefore of defined structure and composition. These unique safety properties of *S/MARs* make them very attractive for use in gene therapy with large genomic constructs.
