**3. Delivery of large DNA molecules**

Gene therapy with large genomic constructs cloned on BAC vectors unavoidably raises the issue of delivery into target cells. The majority of recombinant viruses, commonly utilized as carriers for transfer of plasmid DNA, apart from evoking unwanted immune responses, have a maximum packaging capacity and cannot be used to deliver large BACs. An exception to this rule is the *Herpes simplex* virus 1 (HSV-1) -derived amplicon vector, which has been shown to be able to accommodate and deliver large BACs of up to approximately 150 kb in size (Wade-Martins et al., 2003). However, even this promising vector is based on viral sequences and is subject to criticism regarding its safety. Therefore, delivery of genomic loci of therapeutic genes should preferably be non-viral.

Irrespective of their size, naked DNA molecules are difficult to transfect into cells both *in vivo* and *in vitro* due to a series of barriers related to almost all aspects of cellular biology. Such barriers include degradation by nucleases present in the blood and the extracellular matrix, the plasma membrane, transformation of endosomes to digestive lysozomes following endocytosis and the nuclear envelope (Al-Dosari & Gao, 2009). Several physical methods have been employed to facilitate transfer of naked DNA into cells with efficiencies that in some cases resemble those of viral methods. In parallel, chemical methods based mostly on cationic lipids and polymers have been developed and used for *in vitro* and *in vivo* gene transfer. However, the vast majority of published data concern delivery of plasmid DNA and further evaluation of both physical and chemical methods for the delivery of large BACs is required.

### **3.1 Physical methods for direct delivery to patients**

Physical methods facilitate entry of naked DNA into the cells by creating temporary microdisruption of the cell membrane due to physical forces, such as hydrostatic pressure, electric pulse, ultrasound, laser irradiation, magnetic fields and particle bombardment. As of March 2011, naked plasmid DNA has been used in 18.7% (n=319) of clinical gene therapy trials (http://www.wiley.com/legacy/wileychi/genmed/clinical/).

#### **3.1.1 Injection**

4 Non-Viral Gene Therapy

Transgenic work has also been carried out with respect to cystic fibrosis (CF). The disease is caused by mutations in the cystic fibrosis transmembrane conductance regulator (*CFTR*) which is normally expressed in specific tissues and shows precisely regulated expression thanks to some DHS that are found as far as 80 kb upstream of the first exon (Smith, et al., 1995). Previous experiments with small mini-gene constructs that obviously could not cover the whole genomic region showed some expression of *CFTR* in transgenic mice (Alton et al., 1993; Hyde et al., 1993) and low levels of transient correction in CF patients (Caplen et al., 1995). However, such constructs were not expressed sufficiently in the appropriate tissues to achieve clinical improvement in patients. A 320-kb YAC carrying the human *CFTR* gene rescued CFTR null mice (Manson et al., 1997). Gene expression followed the wild type mouse pattern except in some tissues such as the pyloric glands, Brunner's glands, epididymis and sublingual glands, presumably due to absence of a distant DHS in the YAC or lack of recognition of the human control elements by mouse transcription factors. More recently, correct temporal and spatial expression of the human *COL6A1* gene (Xing et al., 2007), the human Brain-Derived neurotropic factor (*BDNF*) gene (Koppel et al., 2009) and the porcine Growth Hormone gene (Tong et al., 2010) has been demonstrated in BAC

The ability of genomic constructs to drive tissue- and time-specific expression, unlike cDNA and mini-genes, has made BAC transgenics an additional tool to knockout transgenics for the identification of potential regulatory elements within the locus of a gene. For instance, an enhancer within the locus of the tyrosinase-related family (*Tyrp1*) gene (Murisier et al., 2006) and a region responsible for tissue-specific expression within the locus of the *Neurogenin1* gene (Quiñones et al., 2010) have been discovered in transgenic mice generated

Gene therapy with large genomic constructs cloned on BAC vectors unavoidably raises the issue of delivery into target cells. The majority of recombinant viruses, commonly utilized as carriers for transfer of plasmid DNA, apart from evoking unwanted immune responses, have a maximum packaging capacity and cannot be used to deliver large BACs. An exception to this rule is the *Herpes simplex* virus 1 (HSV-1) -derived amplicon vector, which has been shown to be able to accommodate and deliver large BACs of up to approximately 150 kb in size (Wade-Martins et al., 2003). However, even this promising vector is based on viral sequences and is subject to criticism regarding its safety. Therefore, delivery of

Irrespective of their size, naked DNA molecules are difficult to transfect into cells both *in vivo* and *in vitro* due to a series of barriers related to almost all aspects of cellular biology. Such barriers include degradation by nucleases present in the blood and the extracellular matrix, the plasma membrane, transformation of endosomes to digestive lysozomes following endocytosis and the nuclear envelope (Al-Dosari & Gao, 2009). Several physical methods have been employed to facilitate transfer of naked DNA into cells with efficiencies that in some cases resemble those of viral methods. In parallel, chemical methods based mostly on cationic lipids and polymers have been developed and used for *in vitro* and *in vivo* gene transfer. However, the vast majority of published data concern delivery of plasmid DNA and further evaluation of both physical and chemical methods for the delivery of large

transgenic mice.

with BACs.

BACs is required.

**3. Delivery of large DNA molecules** 

genomic loci of therapeutic genes should preferably be non-viral.

In early experiments, naked plasmid DNA (or RNA) was injected directly into different organs of different animals and expression of reporter genes was detected at the sites of injection. In mice, this method was used to deliver and express reporter genes in several organs including skeletal muscle (Wolff et al., 1990) and lung (Meyer et al., 1995). Subsequently, expression of therapeutic genes has been achieved using this method and some human clinical trials for limb ischemia (Morishita et al., 2010), erectile dysfunction (Melman et al., 2007), Duchenne/Becker muscular dystrophy (Romero et al., 2004)) have been based on naked DNA injection into tissues. Although these were phase I or IIa trials aiming mostly at assessing safety, some transgene expression and clinical improvement has been shown but the level of expression was low.

A non-invasive alternative to conventional needle injection is jet injection. This technology is based on a high velocity narrow jet of liquid containing the DNA, which is able to penetrate the skin and underlying tissues. It is powered by compressed air and penetration in a specific tissue can be controlled by adjusting the air pressure. So far, the major application of jet injection has been the development of DNA vaccines (Raviprakash & Porter, 2006). Thanks to technical improvements, the efficiency of delivery by jet injection has reached that of other non-viral methods and has been evaluated recently in a clinical trial on patients with melanoma and breast cancer (Walther et al., 2008).

Further progress in the field has led to the development of the so-called hydrodynamic injection, which is considered to be the most efficient non-viral gene delivery method in mice (Al-Dosari & Gao, 2009). According to the hydrodynamic method, a high volume saline solution of plasmid DNA is injected into the tail vein at high velocity. Initial studies have shown that this results in high levels of gene expression in the liver (Liu, et al., 1999; Zhang et al., 1999). The hydrodynamic injection into the tail vain has also been shown to work relatively well with large BAC DNA (Hibbitt et al., 2007; Magin-Lachmann et al., 2004). Moreover, local hydrodynamic delivery into rabbit liver using catheter-assisted perfusion (Eastman et al., 2002) and pressure-mediated delivery to rat kidney (Maruyama et al., 2002) and to limb muscle of mammals (Hagstrom et al., 2004) have been achieved. it remains to be seen whether similar hydrodynamic methods could be deployed in human patients.

#### **3.1.2 Electroporation**

*In vivo* electroporation is the application of electrical pulses following local injection of DNA in the target tissue. This temporarily increases the cell membrane permeability and facilitates DNA uptake by a mechanism that remains unclear. Under optimal conditions the efficiency of plasmid DNA delivery by *in vivo* electroporation can approach that of viral methods but the efficiency decreases when larger DNA molecules are used. By using a variety of electrodes, ranging from needle to surface electrodes, electroporation has been shown to be effective at

Non-Viral Gene Therapy Vectors Carrying Genomic Constructs 7

Finally, magnetic fields have been used to enhance *in vivo* targeted gene delivery. In this method, called magnetofection, the DNA is reversibly coupled to superparamagnetic nanoparticles which are directed to the target site, following local injection, via a highenergy magnetic field. *In vivo* magnetofection has been shown to work for small in size plasmid DNA delivery to the gastrointestinal tract of rats and the blood vessels of the ear of pigs (Scherer et al., 2002) and the respiratory epithelium of mice (Xenariou et al., 2006).

Chemical vectors used for gene delivery present a broad diversity, with hundreds of different reagents being available (Al-Dosari & Gao, 2009), but generally fall into two main categories: cationic lipids and cationic polymers. These act by forming complexes with the negatively charged DNA, named lipoplexes and polyplexes respectively, which protect the DNA from nucleases and allow its entry into the cells by endocytosis, pinocytarosis or phagocytosis and its transfer into the nucleus by escaping of the complexes from the endosomes following their internalization. The mechanism by which these processes take place are different for lipoplexes and polyplexes and the overall efficiency depends on the chemical structure of the cationic lipids/polymers, the charge ratio between the cationic lipids/polymers and the DNA, the size and structure of the lipoplexes/polyplexes and the inclusion of helper lipids such as DOPE and cholesterol in the complexes (Tros de

Over the years many cationic lipids showing high transfection efficiency were developed and lipofection has become the most common method for gene transfer *in vitro*. Unfortunately, lipoplexes are not equally good for *in vivo* delivery as most of them are inactivated after interaction with factors present in the blood. However, successful *in vivo* DNA delivery has been reported both with systemic and local administration. A single intravenous injection of lipoplexes into mice has been shown to result in reporter transgene expression in the lung, heart, liver, spleen, and kidney (Song et al., 1997). Impressively, local administration of cationic lipid/*CFTR*-plasmid-DNA complexes in an aerosol formulation to the lungs of cystic fibrosis transgenic mice resulted in correction of the ion transport defect (Hyde et al., 1993). Similar studies in human patients demonstrated some transgene expression but not at sufficient levels to provide a clinical benefit (Griesenbach and

Apparently, the choice of the cationic lipid to be used depends on the application and careful optimization of the transfection protocol is required considering that lipoplexes can also induce unwanted immune responses. Lipofection has been used in 6.4% (n=109) of gene

There has also been an extensive use of cationic polymers in DNA delivery studies with polyethylenimine (PEI) being the most active reagent. Polyplexes are more stable than lipoplexes and their toxicity and transfection activity depend on their molecular weigh (mw). Polymers with low mw are more efficient and less toxic than those with high mw (Fischer et al., 1999). Interestingly, intravenous injection of PEI/BAC-DNA complexes in mice has been found to be less efficient than other non-viral gene delivery methods such as

An alternative to *in vivo* delivery of DNA for gene therapy is the *ex vivo* approach. This procedure consists of surgically removing target cells from a patient, transducing them with

therapy clinical trials (http://www.wiley.com/legacy/wileychi/genmed/clinical/).

electroporation and hydrodynamic injection (Magin-Lachmann et al., 2004).

**3.2 Chemical methods for direct delivery to patients** 

Ilarduya et al., 2010).

Alton, 2009).

**3.3** *Ex vivo* **delivery to cells** 

delivering small in size DNA to several tissues including muscle and lung (Brown et al., 2008; Dean et al., 2003). On the other hand, only few data is available about *in vivo* electroporation of large DNA molecules. In one study, efficient delivery of an 80-kb BAC into electroporated muscle has been achieved but, as expected, reporter gene expression from the BAC was found 5-fold less efficient than from a plasmid (Magin-Lachmann et al., 2004). Increasing knowledge and technological progress in electroporation has resulted in its clinical application in humans for the treatment of melanoma (Daud et al., 2008) and in several ongoing clinical trials for the treatment of other cancers and for DNA vaccination.

#### **3.1.3 Sonoporation**

Sonoporation is a technique that uses ultrasound waves of high intensity and low frequency to cause the same effect on the plasma membrane as electroporation that is transient permeabilization in order to facilitate the delivery of DNA into cells. The mechanism is different to electroporation though. Ultrasound is believed to result in acoustic cavitation that can disrupt temporarily the cell membrane. When it is used in combination with microbubbles, which are gas-filled vesicles coated with albumin, polymers or phospholipids, cavitation and therefore local DNA uptake are enhanced (Wells, 2004). Such microbubbles are commercially available and their stability has been shown to affect directly the efficiency of *in vivo* sonoporation (Alter et al., 2009).

Several studies have shown *in vivo* delivery of plasmids carrying either reporter or therapeutic genes to different tissues including lung, heart and muscle (Xenariou et al., 2010; Alter et al., 2009; Sheyn et al., 2008) but comparative data, wherever provided, confirmed that the efficiency of sonoporation was significantly lower than that of electroporation. However, sonoporation is still being considered for clinical application in humans due to its non-invasive nature and lesser tissue damage caused compared to electroporation. Interestingly, a combination of electric pulses and ultrasound waves (electrosonoporation) for gene transfer into the skeletal muscle of mice showed 8-fold and 1.6-fold higher gene expression compared with electroporation and sonoporation alone, respectively (Yamashita et al., 2002).

#### **3.1.4 Other physical methods**

Particle bombardment via a gene gun, originally designed for DNA delivery in plants, is a non-viral gene transfer method based on gold particles coated with DNA. The particles are accelerated by pressurized gas and expelled onto tissues. This technique, also referred to as ballistic DNA delivery, has been used to deliver transgenes to skin, liver and muscle tissues of rats and mice (Yang et al., 1990) and DNA vaccines to skin, muscles and tumours in animal models and in human clinical trials. Recently, is has been shown to be efficient at delivering a small reporter gene to mouse liver *in vivo* (Chang et al., 2008). Almost no data is available on the ability of gene gun to deliver large DNA constructs *in vivo*. In just one study, a DNA vaccine containing a 183-kb BAC has been delivered using gene gun and has been shown to confer immune protection to chickens (Tischer et al., 2002).

A new promising method of gene transfer based on the utilization of infrared laser beam has been developed and used to deliver a small transgene to mouse muscle *in vivo* (Zeira et al., 2003). This study reported efficiency of delivery, assessed by measuring the intensity and duration of transgene expression, equal to that by electroporation but with less damage caused to the tissue.

delivering small in size DNA to several tissues including muscle and lung (Brown et al., 2008; Dean et al., 2003). On the other hand, only few data is available about *in vivo* electroporation of large DNA molecules. In one study, efficient delivery of an 80-kb BAC into electroporated muscle has been achieved but, as expected, reporter gene expression from the BAC was found 5-fold less efficient than from a plasmid (Magin-Lachmann et al., 2004). Increasing knowledge and technological progress in electroporation has resulted in its clinical application in humans for the treatment of melanoma (Daud et al., 2008) and in several ongoing clinical trials for the

Sonoporation is a technique that uses ultrasound waves of high intensity and low frequency to cause the same effect on the plasma membrane as electroporation that is transient permeabilization in order to facilitate the delivery of DNA into cells. The mechanism is different to electroporation though. Ultrasound is believed to result in acoustic cavitation that can disrupt temporarily the cell membrane. When it is used in combination with microbubbles, which are gas-filled vesicles coated with albumin, polymers or phospholipids, cavitation and therefore local DNA uptake are enhanced (Wells, 2004). Such microbubbles are commercially available and their stability has been shown to affect directly

Several studies have shown *in vivo* delivery of plasmids carrying either reporter or therapeutic genes to different tissues including lung, heart and muscle (Xenariou et al., 2010; Alter et al., 2009; Sheyn et al., 2008) but comparative data, wherever provided, confirmed that the efficiency of sonoporation was significantly lower than that of electroporation. However, sonoporation is still being considered for clinical application in humans due to its non-invasive nature and lesser tissue damage caused compared to electroporation. Interestingly, a combination of electric pulses and ultrasound waves (electrosonoporation) for gene transfer into the skeletal muscle of mice showed 8-fold and 1.6-fold higher gene expression compared with electroporation and sonoporation alone, respectively (Yamashita

Particle bombardment via a gene gun, originally designed for DNA delivery in plants, is a non-viral gene transfer method based on gold particles coated with DNA. The particles are accelerated by pressurized gas and expelled onto tissues. This technique, also referred to as ballistic DNA delivery, has been used to deliver transgenes to skin, liver and muscle tissues of rats and mice (Yang et al., 1990) and DNA vaccines to skin, muscles and tumours in animal models and in human clinical trials. Recently, is has been shown to be efficient at delivering a small reporter gene to mouse liver *in vivo* (Chang et al., 2008). Almost no data is available on the ability of gene gun to deliver large DNA constructs *in vivo*. In just one study, a DNA vaccine containing a 183-kb BAC has been delivered using gene gun and has

A new promising method of gene transfer based on the utilization of infrared laser beam has been developed and used to deliver a small transgene to mouse muscle *in vivo* (Zeira et al., 2003). This study reported efficiency of delivery, assessed by measuring the intensity and duration of transgene expression, equal to that by electroporation but with less damage

been shown to confer immune protection to chickens (Tischer et al., 2002).

treatment of other cancers and for DNA vaccination.

the efficiency of *in vivo* sonoporation (Alter et al., 2009).

**3.1.3 Sonoporation** 

et al., 2002).

caused to the tissue.

**3.1.4 Other physical methods** 

Finally, magnetic fields have been used to enhance *in vivo* targeted gene delivery. In this method, called magnetofection, the DNA is reversibly coupled to superparamagnetic nanoparticles which are directed to the target site, following local injection, via a highenergy magnetic field. *In vivo* magnetofection has been shown to work for small in size plasmid DNA delivery to the gastrointestinal tract of rats and the blood vessels of the ear of pigs (Scherer et al., 2002) and the respiratory epithelium of mice (Xenariou et al., 2006).

#### **3.2 Chemical methods for direct delivery to patients**

Chemical vectors used for gene delivery present a broad diversity, with hundreds of different reagents being available (Al-Dosari & Gao, 2009), but generally fall into two main categories: cationic lipids and cationic polymers. These act by forming complexes with the negatively charged DNA, named lipoplexes and polyplexes respectively, which protect the DNA from nucleases and allow its entry into the cells by endocytosis, pinocytarosis or phagocytosis and its transfer into the nucleus by escaping of the complexes from the endosomes following their internalization. The mechanism by which these processes take place are different for lipoplexes and polyplexes and the overall efficiency depends on the chemical structure of the cationic lipids/polymers, the charge ratio between the cationic lipids/polymers and the DNA, the size and structure of the lipoplexes/polyplexes and the inclusion of helper lipids such as DOPE and cholesterol in the complexes (Tros de Ilarduya et al., 2010).

Over the years many cationic lipids showing high transfection efficiency were developed and lipofection has become the most common method for gene transfer *in vitro*. Unfortunately, lipoplexes are not equally good for *in vivo* delivery as most of them are inactivated after interaction with factors present in the blood. However, successful *in vivo* DNA delivery has been reported both with systemic and local administration. A single intravenous injection of lipoplexes into mice has been shown to result in reporter transgene expression in the lung, heart, liver, spleen, and kidney (Song et al., 1997). Impressively, local administration of cationic lipid/*CFTR*-plasmid-DNA complexes in an aerosol formulation to the lungs of cystic fibrosis transgenic mice resulted in correction of the ion transport defect (Hyde et al., 1993). Similar studies in human patients demonstrated some transgene expression but not at sufficient levels to provide a clinical benefit (Griesenbach and Alton, 2009).

Apparently, the choice of the cationic lipid to be used depends on the application and careful optimization of the transfection protocol is required considering that lipoplexes can also induce unwanted immune responses. Lipofection has been used in 6.4% (n=109) of gene therapy clinical trials (http://www.wiley.com/legacy/wileychi/genmed/clinical/).

There has also been an extensive use of cationic polymers in DNA delivery studies with polyethylenimine (PEI) being the most active reagent. Polyplexes are more stable than lipoplexes and their toxicity and transfection activity depend on their molecular weigh (mw). Polymers with low mw are more efficient and less toxic than those with high mw (Fischer et al., 1999). Interestingly, intravenous injection of PEI/BAC-DNA complexes in mice has been found to be less efficient than other non-viral gene delivery methods such as electroporation and hydrodynamic injection (Magin-Lachmann et al., 2004).

#### **3.3** *Ex vivo* **delivery to cells**

An alternative to *in vivo* delivery of DNA for gene therapy is the *ex vivo* approach. This procedure consists of surgically removing target cells from a patient, transducing them with

Non-Viral Gene Therapy Vectors Carrying Genomic Constructs 9

invasin, from the *inv* gene of *Yersinia pseudotuberculosis*, which binds to β1-integrins on mammalian cells leading to internalization, 2) have permanent impaired cell wall synthesis due to diaminopimelic acid (DAP) auxotrophy which causes bacterial lysis following internalization and 3) stably express listeriolysin O, from the *hly* locus of *Listeria monocytogenes*, which is a pore-forming cytolysin that allows escape from the vacuole after bacterial entry and release of the DNA into the cytosol resulting in greater levels of

One advantage of the method is that there is no need to purify the DNA prior to transfection, which is still a technically challenging procedure for very large DNA molecules. Another advantage is that DNA delivered by bacterial invasion is rarely rearranged. Rearrangements are always an issue of concern when using other DNA delivery methods to transfer large BACs. For instance, although lipofection and electroporation have been used to efficiently deliver large BAC DNA into cultured cells, some of the clones generated have been shown to suffer from rearrangements (Magin-Lachmann et al., 2004; Cheung et al., unpublished data). In contrast, several studies have demonstrated that large BACs are delivered intact by bacterial invasion albeit with low efficiency. In two of them, stable cell lines containing integrations of a ~250-kb BAC carrying the human clotting factor VIII *(FVIII)* gene (Pérez-Luz et al., 2007) and of a ~258-kb BAC carrying the human *CFTR* gene (Kotzamanis et al., 2009) have been generated by bacterial invasion. No clones have been found to contain rearranged DNA in any of the two studies and expression of the respective transgenes has also been shown in both of them. Therefore, bactofection is an attractive method for delivering large genomic DNA containing intact constructs. The ability of the method to deliver large BACs into MSCs without affecting their differentiation

Regardless of the type and size of the therapeutic gene (small cDNA versus large genomic DNA) and the delivery method (viral versus non-viral) to be used in a gene therapy protocol, efficient retention and long-term expression of the transgene is required so as to eliminate the need for re-administrations. Integration into the host genome has widely been used in gene therapy to fulfil this requirement. However, the dangers of integration due to insertional mutagenesis have become a widely publicised issue as a result of a clinical trial using a retroviral vector to treat X-linked severe combined immune deficiency (SCIDX1). In this trial some patients developed leukaemia due to deregulation of the growth-promoting LIM domain only 2 (*LMO2*) proto-oncogene caused by integration of the vector (Hacein-Bey-Abina et al., 2003a, 2003b). The safety concerns regarding uncontrolled integration of the therapeutic gene into the host genome have been strengthened by observations that there is a preference of integrating vectors for the regulatory regions of transcriptionally active genes (Bushman et al., 2005). Given the need for long-term expression and the problems associated with vector integration, vectors that persist in the nucleus by being maintained episomally without integrating could be highly advantageous. Three different systems have been employed to achieve extra-chromosomal maintenance of the vectors carrying the therapeutic gene: systems based on elements from the Epstein-Barr virus (EBV), artificial chromosomes and systems based on *scaffold/matrix attachment region* (*S/MAR*). All these systems have a high cloning capacity and can be used in combination with large

transgene expression (Laner et al., 2005).

potential is yet to be shown.

genomic constructs.

**4. Extra-chromosomal vectors** 

an appropriate therapeutic gene in culture and then reimplanting them into the body of the donor. Since it involves the use of autologous cells, there is no need for immunosupression following transplantation of the cells back to the patient. In most *ex vivo* gene therapy applications, transplanted cells need to integrate into the appropriate tissue and the efficiency of this "homing" process is tissue type-dependent.

Recent progress in stem cell research has revolutionized *ex vivo* therapy. Stem cells are characterized by their ability to differentiate into a diverse range of cell types when placed in the appropriate environment both *in vitro* and *in vivo* and can therefore be used directly in cell therapy (Abdulrazzak et al., 2010). An *ex vivo* cell therapy approach using Embryonic Stem Cells or foetal Mesenchymal Stem Cells would not involve transfection of a therapeutic gene and would overcome the low gene delivery efficiency hurdle. However, it would have to be used in an allogeneic fashion which would bring the need for using immunosuppressive drugs. A more attractive strategy would be the transfer of a therapeutic gene to patient-derived autologous stem cells such as Mesenchymal Stem Cells (MSCs) which can easily be isolated from the bone marrow or adipose tissue of adults and have been shown to have an excellent differentiation potential (Abdallah & Kassem, 2008) or Induced Pluripotent Stem Cells that can be generated by reprogramming of adult somatic cells (Yu et al., 2007; Takahashi et al., 2007).

Several non-viral methods have been utilized to transfer DNA into MSCs for *ex vivo* gene therapy purposes. The aim of the following sections is to review two of them, nucleofection and bactofection. The criteria for this choice are their ability to deliver very large DNA molecules intact and/or maintain the differentiation potential of the MSCs following transfection.

#### **3.3.1 Nucleofection**

Nucleofector technology developed by Amaxa Biosystems is a non-viral method of gene transfer based on electroporation using a combination of electrical parameters and solutions that are specific for each cell type. Unlike other non-viral transfection methods, it facilitates the transfer of DNA directly into the cell nucleus. It is particularly useful for gene transfer into a variety of primary cell types from different species which are normally very difficult to transfect (Gresch et al., 2004).

 Nucleofection has been shown to be able to efficiently deliver plasmid DNA carrying a reporter gene to MSCs without impairment of their differentiation capacity (Aluigi et al., 2006). Preliminary data on stable cell lines generated by nucleofection with large BACs suggest that nucleofection can also be used to deliver genomic constructs but does not result in all clones containing intact and unrearranged DNA (Cheung et. al., unpublished data).

#### **3.3.2 Bactofection**

Direct delivery of DNA into mammalian cells by invasive bacteria (bactofection) is another potentially useful technique for gene transfer and it may have applications for both *in vitro* and *in vivo* delivery (Larsen et al., 2008). In this method, the DNA is first introduced either in the form of a plasmid or a BAC into bacteria having the ability to invade eukayotic cells and subsequently these bacteria are used to invade and deliver their DNA content into target cells. Several bacterial systems allowing eukaryotic cell invasion have been described. The most convenient is based on the the *E. coli* strain BM4573 (Laner et al., 2005). BM4573 bacteria have been modified to 1) stably express

an appropriate therapeutic gene in culture and then reimplanting them into the body of the donor. Since it involves the use of autologous cells, there is no need for immunosupression following transplantation of the cells back to the patient. In most *ex vivo* gene therapy applications, transplanted cells need to integrate into the appropriate tissue and the

Recent progress in stem cell research has revolutionized *ex vivo* therapy. Stem cells are characterized by their ability to differentiate into a diverse range of cell types when placed in the appropriate environment both *in vitro* and *in vivo* and can therefore be used directly in cell therapy (Abdulrazzak et al., 2010). An *ex vivo* cell therapy approach using Embryonic Stem Cells or foetal Mesenchymal Stem Cells would not involve transfection of a therapeutic gene and would overcome the low gene delivery efficiency hurdle. However, it would have to be used in an allogeneic fashion which would bring the need for using immunosuppressive drugs. A more attractive strategy would be the transfer of a therapeutic gene to patient-derived autologous stem cells such as Mesenchymal Stem Cells (MSCs) which can easily be isolated from the bone marrow or adipose tissue of adults and have been shown to have an excellent differentiation potential (Abdallah & Kassem, 2008) or Induced Pluripotent Stem Cells that can be generated by reprogramming of adult somatic

Several non-viral methods have been utilized to transfer DNA into MSCs for *ex vivo* gene therapy purposes. The aim of the following sections is to review two of them, nucleofection and bactofection. The criteria for this choice are their ability to deliver very large DNA molecules intact and/or maintain the differentiation potential of the MSCs following

Nucleofector technology developed by Amaxa Biosystems is a non-viral method of gene transfer based on electroporation using a combination of electrical parameters and solutions that are specific for each cell type. Unlike other non-viral transfection methods, it facilitates the transfer of DNA directly into the cell nucleus. It is particularly useful for gene transfer into a variety of primary cell types from different species which are normally very difficult

 Nucleofection has been shown to be able to efficiently deliver plasmid DNA carrying a reporter gene to MSCs without impairment of their differentiation capacity (Aluigi et al., 2006). Preliminary data on stable cell lines generated by nucleofection with large BACs suggest that nucleofection can also be used to deliver genomic constructs but does not result in all clones containing intact and unrearranged DNA (Cheung et. al., unpublished data).

Direct delivery of DNA into mammalian cells by invasive bacteria (bactofection) is another potentially useful technique for gene transfer and it may have applications for both *in vitro* and *in vivo* delivery (Larsen et al., 2008). In this method, the DNA is first introduced either in the form of a plasmid or a BAC into bacteria having the ability to invade eukayotic cells and subsequently these bacteria are used to invade and deliver their DNA content into target cells. Several bacterial systems allowing eukaryotic cell invasion have been described. The most convenient is based on the the *E. coli* strain BM4573 (Laner et al., 2005). BM4573 bacteria have been modified to 1) stably express

efficiency of this "homing" process is tissue type-dependent.

cells (Yu et al., 2007; Takahashi et al., 2007).

transfection.

**3.3.1 Nucleofection** 

**3.3.2 Bactofection** 

to transfect (Gresch et al., 2004).

invasin, from the *inv* gene of *Yersinia pseudotuberculosis*, which binds to β1-integrins on mammalian cells leading to internalization, 2) have permanent impaired cell wall synthesis due to diaminopimelic acid (DAP) auxotrophy which causes bacterial lysis following internalization and 3) stably express listeriolysin O, from the *hly* locus of *Listeria monocytogenes*, which is a pore-forming cytolysin that allows escape from the vacuole after bacterial entry and release of the DNA into the cytosol resulting in greater levels of transgene expression (Laner et al., 2005).

One advantage of the method is that there is no need to purify the DNA prior to transfection, which is still a technically challenging procedure for very large DNA molecules. Another advantage is that DNA delivered by bacterial invasion is rarely rearranged. Rearrangements are always an issue of concern when using other DNA delivery methods to transfer large BACs. For instance, although lipofection and electroporation have been used to efficiently deliver large BAC DNA into cultured cells, some of the clones generated have been shown to suffer from rearrangements (Magin-Lachmann et al., 2004; Cheung et al., unpublished data). In contrast, several studies have demonstrated that large BACs are delivered intact by bacterial invasion albeit with low efficiency. In two of them, stable cell lines containing integrations of a ~250-kb BAC carrying the human clotting factor VIII *(FVIII)* gene (Pérez-Luz et al., 2007) and of a ~258-kb BAC carrying the human *CFTR* gene (Kotzamanis et al., 2009) have been generated by bacterial invasion. No clones have been found to contain rearranged DNA in any of the two studies and expression of the respective transgenes has also been shown in both of them. Therefore, bactofection is an attractive method for delivering large genomic DNA containing intact constructs. The ability of the method to deliver large BACs into MSCs without affecting their differentiation potential is yet to be shown.
