**9. rAAV and gene editing technology**

are 1% of normal levels will bleed into the joint and muscle tissues. Bleeding in the brain could result in fatal death. If FIX levels are between 1% and 5%, the individual will experience a reduced number of bleeding incidents and a moderate phenotype of the disease. Any FIX levels above 5% will allow the person to have a normal life [53]. The only available treatment is protein replacement therapy, which requires regular intravenous injections and is expensive. Therefore, novel and permanent therapies/treatments are urgent. rAAV gene therapy cur‐ rently constitutes a promising approach for the treatment of several diseases, including

Based on animal studies that were described in the previous section, four clinical trials have been initiated. The first study administered rAAV carrying FIX gene into three patients by intramuscular injection. Despite the presence of preexistent high titer of neutralizing antibod‐ ies against capsid rAAV, strong transgene expression was observed in the muscle, even after 10 months of injection. However, levels of factor IX in circulation were less than 1–2% in most

The second trial conducted by the University of Pennsylvania infused the virus through the hepatic artery into seven patients. The rational in this protocol considered that FIX is a secreted protein and once it is produced and reaches the bloodstream; it can be distributed throughout

Even though levels of FIX resulted higher than 5% after injecting an intermediate or high viral dose, the therapeutic effect was only transient (up to eight weeks), due to the development of a strong cytotoxic T response, which destroyed the transfected hepatocytes and thus hampered

The third trial was designed in order to increase FIX expression production as well as to circumvent the possibility of a humoral response that could interfere with the success of the therapy [57]. To reach the first goal, they developed a codon optimized FIX gene that also delivered the gene in the context of self-complementary rAAV, which provides substantially higher levels of transgene expression rather than delivering the WT gene with single-stranded rAAV. In order to reduce antibody neutralization, the viral genome cassette was packed in rAAV8 capsid, as it has lower seroprevalence in humans and a high tropism for the liver. The virus was administered directly in the peripheral vein in six patients, and all of them developed 1–6% levels of factor IX expression in the first four months as well as for at least three years. There was no modification on the levels of neutralizing antibodies. However, transient elevations in serum liver enzymes, possibly as a result of a cellular immune response to the

rAAV8 capsid, were observed in the three patients who received the high viral dose.

Recently, Baxter's laboratories launched a clinical trial to test the safety and efficacy of a selfcomplimentary rAAV8 vector carrying a mutant FIX sequence (BAX 335), created and preclinically tested at the UNC gene therapy center [58]. The FIX sequence had a single amino acid change at position 338, which substantially increased the levels of circulating FIX protein. A more effective rAAV delivery vehicle allowed the administration of lower viral doses with the same efficacy as previous tested virus, but without the associated toxicity. In this trial, sixteen adult patients were treated in different centers around the US. Updates on the trial

cases, even at the highest tested dose. Toxicity was not observed [54, 55].

hemophilia B.

134 Gene Therapy - Principles and Challenges

the body.

the production of FIX [56].

Several genome editing tools have emerged recently in an attempt to correct the genetic cause(s) of a disease. These technologies rely on two components: a sequence-specific DNAbinding domain and a nuclease [59]. The procedure consists of several steps: (1) recognition of a targeted DNA sequence, (2) double stranded cut and (3) stimulation of a cellular repair mechanism to correct the DNA damage, which includes homologous recombination [60].

This technology allows for modifying a coding sequence, the epigenome, transcriptional activator/repressor as well as a regulatory element such as transcription factors, recombinases, transposases, and more. When targeting a particular gene, these technologies generate deletions, insertions or mutations of the gene, which may be useful to elucidate the gene function, or to generate cell lines with the null phenotype, or even to model a specific genetic condition for its study. Three different systems are currently available: zinc-finger nucleases (ZFN), TALEN nucleases and CRISPR/Cas9 [60].

Zinc-finger nucleases are a common type of DNA-binding motif found in eukaryotes and therefore, in the human genome. Usually, the DNA binding domain in the zinc-finger nuclease recognizes three base pairs in the DNA sequence. However, researchers have engineered the domain in order to detect and bind any defined DNA sequence of 9 to 18 bps in length, allowing the targeting of up to 68 billion bp of DNA [61].

Even though these technologies are very promising, an optimal delivery vehicle of the gene editing system needs to be developed. rAAV has the potential to deliver nucleases in vitro and in vivo and also has the potential to induce homologous recombination in the cell that infects, further enhancing the homologous recombination efficiency by 1000-fold [62–64].

In 2011, High et al. showed the generation of a ZFN system capable of cleaving F9 intron 1 and inducing homology-directed repair in the human hepatocyte Hep3B cell line. As proof of principle, the system induced up to 17% stable integration of a novel restriction enzyme site into the F9 locus. Furthermore, intraperitoneal administration of a ZFN system, which specifically targets F9, via rAAV8 delivery, in conjunction with an rAAV8 vector carrying a corrective F9 complementary DNA cassette into a humanized neonatal mouse model of hemophilia B, resulted in 1%–3% specific targeting of mouse liver. However, this mouse at two days old produced 2–3% normal F9 levels, enough amount to convert severe to mild hemophilia [65]. In 2013, they tested the same technology in a young adult mouse (8–10 weeks old), in which hepatocyte proliferation is slow as the liver already reached its maturity [66]. In theory, younger mice should show higher levels of gene correction, compared to older mice, as mice age affects the rate of homologous recombination, which is essential for genome editing to occur. In this publication, even though adult mice showed limited hepatocyte proliferation, following AAV injection, mice experienced a 5-fold increase in FIX expression, compared to the previous study. Moreover, when they tested the technology in even older mice, 7–8 months old, FIX levels were extremely low. Investigators argue that the discrepancies in FIX levels between neonate and adult mice could be attributable to the loss of rAAV vector genomes during liver development and/or different promoter activity. Furthermore, when they switched the use of homodimer nucleases to heterodimeric ZFN, nonespecific ZFN cleavage was observed without the loss of FIX expression.

Additionally, ZFN technology is currently being investigated in clinical trials for the treatment of HIV. Basically, the therapy consists of ex vivo permanent modification of patient T cells to knock down the HIV entry receptor CCR5 and autologous administration of the recombinant cells back to the patient. This clinical trial is sponsored by Sangamo Biosciences, the same company that collaborated with D. High for the in vivo targeting of hemophilia B mouse with rAAV-ZFN platform. In 2014, the company released an announcement for the first IND to test ZFN genome-editing platform in hemophilia A patients.

Even though these studies performed by Dr. High's laboratory and Sangamo Biosciences showed potential for in vivo gene editing via rAAV delivery, especially for diseases which do not allow ex vivo manipulation of target cells such as hemophilia B, the technology has several issues to address before being considered efficient and safe for treating human patients. First of all, we should consider all the challenges associated with rAAV delivery in vivo, such as the development of a cytotoxic T cell and/or neutralizing antibody responses and exclusively targeting of the tissue to correct with high efficiency. Furthermore, in order for this therapy to work efficiently, each cell needs to receive the two viruses at the same time, a condition that is possible but with a much lower probability to occur, and even if it takes place, the individual would be exposed to higher doses of rAAV8, which enhance the probability of inducing an immune response. Given that homologous recombination repair mainly takes place during the S phase of the cell cycle, gene editing is limited to be successful only in young patients, unless it is combined with molecules/drugs that boost cell division. Importantly, off-target double-stranded DNA breaks pose the possibility of inducing vector integration and/or undesired mutations and consequently, inducing oncogenesis, cell death and/or genetic diseases. This last possible issue could be solved by using self-inactivating viruses. Finally, the ZFN rAAV-mediated technology is still in the early phase of development, so far it has proven its potential for permanently correcting monogenetic diseases. However, considering that (1) rAAV gene therapy has shown great promise in the treatment of hemophilia B and (2) very low levels of FIX are enough to prevent bleeding and allow the person to have a normal life, in vivo gene editing technology seems too risky and unnecessary to pursue for the treatment of hemophilia diseases.

### **10. iPSC and rAAV**

old), in which hepatocyte proliferation is slow as the liver already reached its maturity [66]. In theory, younger mice should show higher levels of gene correction, compared to older mice, as mice age affects the rate of homologous recombination, which is essential for genome editing to occur. In this publication, even though adult mice showed limited hepatocyte proliferation, following AAV injection, mice experienced a 5-fold increase in FIX expression, compared to the previous study. Moreover, when they tested the technology in even older mice, 7–8 months old, FIX levels were extremely low. Investigators argue that the discrepancies in FIX levels between neonate and adult mice could be attributable to the loss of rAAV vector genomes during liver development and/or different promoter activity. Furthermore, when they switched the use of homodimer nucleases to heterodimeric ZFN, nonespecific ZFN cleavage

Additionally, ZFN technology is currently being investigated in clinical trials for the treatment of HIV. Basically, the therapy consists of ex vivo permanent modification of patient T cells to knock down the HIV entry receptor CCR5 and autologous administration of the recombinant cells back to the patient. This clinical trial is sponsored by Sangamo Biosciences, the same company that collaborated with D. High for the in vivo targeting of hemophilia B mouse with rAAV-ZFN platform. In 2014, the company released an announcement for the first IND to test

Even though these studies performed by Dr. High's laboratory and Sangamo Biosciences showed potential for in vivo gene editing via rAAV delivery, especially for diseases which do not allow ex vivo manipulation of target cells such as hemophilia B, the technology has several issues to address before being considered efficient and safe for treating human patients. First of all, we should consider all the challenges associated with rAAV delivery in vivo, such as the development of a cytotoxic T cell and/or neutralizing antibody responses and exclusively targeting of the tissue to correct with high efficiency. Furthermore, in order for this therapy to work efficiently, each cell needs to receive the two viruses at the same time, a condition that is possible but with a much lower probability to occur, and even if it takes place, the individual would be exposed to higher doses of rAAV8, which enhance the probability of inducing an immune response. Given that homologous recombination repair mainly takes place during the S phase of the cell cycle, gene editing is limited to be successful only in young patients, unless it is combined with molecules/drugs that boost cell division. Importantly, off-target double-stranded DNA breaks pose the possibility of inducing vector integration and/or undesired mutations and consequently, inducing oncogenesis, cell death and/or genetic diseases. This last possible issue could be solved by using self-inactivating viruses. Finally, the ZFN rAAV-mediated technology is still in the early phase of development, so far it has proven its potential for permanently correcting monogenetic diseases. However, considering that (1) rAAV gene therapy has shown great promise in the treatment of hemophilia B and (2) very low levels of FIX are enough to prevent bleeding and allow the person to have a normal life, in vivo gene editing technology seems too risky and unnecessary to pursue for the treatment

was observed without the loss of FIX expression.

136 Gene Therapy - Principles and Challenges

ZFN genome-editing platform in hemophilia A patients.

of hemophilia diseases.

iPSC technologies have gained special interest since their discovery in 2006 by Takahashi and Yamanaka [67]. The generation of iPSC has several applications. One of the most important applications consists of the generation of: (1) pluripotent stem cells from a fully differentiated patient cell or (2) a specific human cell that is scarce or not accessible to the scientific population, from a healthy or diseased individual, following differentiation of the pluripotent stem cell. Moreover, sometimes, a personalized treatment is required or a diverse population cell sample is needed for testing the efficacy of a therapeutic technolo‐ gy, such as rAAV. For instance, the common practice is to reprogram patient-derived fibroblasts into a specific cell type that is affected by a disease. Some attempts have been performed to reprogram fibroblasts of patients suffering from retinal diseases into iPSC and finally differentiate the pluripotent stem cells into retinal pigment epithelial (RPE) cells that manifest the diseased phenotype [68, 69]. Following the validation of iPSC and then RPE cells, a panel of rAAV serotypes could be tested for their efficiency to transduce the cells and the most effective ones could be chosen for delivering the healthy gene copy in order to re-establish normal cellular phenotype [70].

Another approach for inducing iPSC development has been tried, but this time, using an rAAV system rather than a retrovirus. IPSC generated by Takahashi and Yamanaka's original protocol made use of retroviruses to deliver Oct3/4, Sox2, Klf4 and c-Myc. Even though the approach resulted in the generation of pluripotent stem cells from mouse embryonic or adult fibroblast cultures, still the efficiency was extremely low, the presence of c-Myc oncogene significantly increased the incidence of tumorigenicity and the use of retrovirus posed the threat of integration into the genome. Several new strategies have been developed, including the use of rAAV [69, 71]. Considering the advantages of using rAAV for gene delivery, such as long-term transgene expression for efficient reprogramming of mature cells as well as safety and efficacy as a gene delivery vehicle in the clinic; research‐ ers have attempted their use in the reprogramming of fully differentiated fibroblasts as well as adipose-derived mesenchymal stem cells. However, both studies observed frequent rAAV integration into the host genome of iPSC cells when the iPSC were generated from nondividing cells. Integration events were independent of the rAAV vector, cell type and amount of virus. Both studies concluded that there is a certain degree of incompatibility between iPSC generation and the use of rAAV vectors, although reprogramming does not require an integration event. Furthermore, like retrovirus-mediated cell reprogramming, rAAV-mediated iPSC generation resulted in reprogramming transgene silencing, which affects the quality of the induced pluripotent stem cells that could be generated. There‐ fore, if the integration events are tightly controlled, which is feasible, and if the epigenet‐ ic mechanisms of rAAV silencing are discovered, rAAV technology could result in a safer mechanism for inducing pluripotent stem cells and consequently, increasing the chances of being applicable to the clinic.
