**5. rAAV targeting the liver for hemophilia treatment**

Even though rAAV therapies for treating hemophilia in mice produced successful results, their translation into large animals, such as a dog or nonhuman primate, was not straightforward [21]. The vector efficacy does not completely follow a dose–response correlation in large animals, and it is drastically affected in the presence of an immunological response towards the viral capsids. Furthermore, in both mice and dogs, there is no direct correlation between the transgene copy numbers and the expression of the foreign protein. However, treatment of a large animal with the therapy was promising as FIX is a secreted protein and only 1%–2% normal factor IX levels is enough to correct the disease [21].

Therapies for hemophilia B were studied using different routes of rAAV gene therapy administration. Initially, intramuscular delivery of rAAV-CMV-cFIX to hemophilic dogs was pursued. A single administration of the virus generated a therapeutic FIX level in a dosedependent manner [22]. However, the amount of antibody formation, and therefore, the success of the therapy, directly correlated with the increase in rAAV dose [23]. Considering that FIX is produced within the liver, delivery of the virus through the portal vein was also attempted. Several steps were performed in order to optimize the vector, such as the addition of a liver specific promoter, and testing different doses to determine the optimum for allowing normal levels of FIX in the dog through this route [24–26]. Mount et al. observed sustained levels of FIX between 4% and 12% at doses between 1.2 × 1012 and 3.4 × 1012 vg/kg for over 17 months in three out of four dogs [26]. The fourth dog experienced a transient correction of FIX levels for four weeks but eventually developed neutralizing antibodies against the transgene. In another study, delivery of a hyperfunctional FIX through rAAV-mediated transfer to the liver in hemophilic dogs resulted in 25% and 300% FIX levels for 1 × 1012 and 3 × 1012 vg/kg doses, respectively [27].

Studies performed with nonhuman primates generated variable outcomes, from no detection to detection up to 10% of FIX in the serum. Failure of the gene therapy was attributed to the preexistence of neutralizing antibodies against the transgene [28, 29]. However, direct administration of the rAAV vector to the liver induced some tolerance to the transgene [27, 28]. Another strategy to overcome the success of the therapy due to the presence of neutralizing antibodies consists of administration of the therapy to an early age of the animal, presumably because the immune system is not completely developed [30].

A comparison study of different administration routes for rAAV-FIX delivery on mice revealed that the same dose allowed 3-fold more transgene expression following intrahepatic rather than intramuscular or intravenous delivery [31]. Therefore, it seems that transgene delivery via rAAV virus is more successful when using the liver-directed gene route. When intrahepatic therapy was administered to dogs who suffered from hemophilia A, rAAV8 carrying the canine factor VII cDNA showed long-term correction of the phenotype, with no spontaneous bleeding episodes, no toxicity and no development of inhibitory antibodies towards the viral vector or the transgene [32]. Similarly, liver-directed rAAV-FIX therapy to dogs suffering from hemophilia B, significantly increased FIX activity to 4%–10% and remained stable for more than eight years [33]. However, direct injection of the virus to the muscle resulted in unde‐ tectable FIX levels in the dog due to the onset of an immune response.

#### **6. Clinical trials using rAAV technology** studies out of 70 were performed in the United States and in 2015, 44 total

Clinical trials using rAAV technology

**5. rAAV targeting the liver for hemophilia treatment**

normal factor IX levels is enough to correct the disease [21].

because the immune system is not completely developed [30].

tectable FIX levels in the dog due to the onset of an immune response.

doses, respectively [27].

128 Gene Therapy - Principles and Challenges

Even though rAAV therapies for treating hemophilia in mice produced successful results, their translation into large animals, such as a dog or nonhuman primate, was not straightforward [21]. The vector efficacy does not completely follow a dose–response correlation in large animals, and it is drastically affected in the presence of an immunological response towards the viral capsids. Furthermore, in both mice and dogs, there is no direct correlation between the transgene copy numbers and the expression of the foreign protein. However, treatment of a large animal with the therapy was promising as FIX is a secreted protein and only 1%–2%

Therapies for hemophilia B were studied using different routes of rAAV gene therapy administration. Initially, intramuscular delivery of rAAV-CMV-cFIX to hemophilic dogs was pursued. A single administration of the virus generated a therapeutic FIX level in a dosedependent manner [22]. However, the amount of antibody formation, and therefore, the success of the therapy, directly correlated with the increase in rAAV dose [23]. Considering that FIX is produced within the liver, delivery of the virus through the portal vein was also attempted. Several steps were performed in order to optimize the vector, such as the addition of a liver specific promoter, and testing different doses to determine the optimum for allowing normal levels of FIX in the dog through this route [24–26]. Mount et al. observed sustained levels of FIX between 4% and 12% at doses between 1.2 × 1012 and 3.4 × 1012 vg/kg for over 17 months in three out of four dogs [26]. The fourth dog experienced a transient correction of FIX levels for four weeks but eventually developed neutralizing antibodies against the transgene. In another study, delivery of a hyperfunctional FIX through rAAV-mediated transfer to the liver in hemophilic dogs resulted in 25% and 300% FIX levels for 1 × 1012 and 3 × 1012 vg/kg

Studies performed with nonhuman primates generated variable outcomes, from no detection to detection up to 10% of FIX in the serum. Failure of the gene therapy was attributed to the preexistence of neutralizing antibodies against the transgene [28, 29]. However, direct administration of the rAAV vector to the liver induced some tolerance to the transgene [27, 28]. Another strategy to overcome the success of the therapy due to the presence of neutralizing antibodies consists of administration of the therapy to an early age of the animal, presumably

A comparison study of different administration routes for rAAV-FIX delivery on mice revealed that the same dose allowed 3-fold more transgene expression following intrahepatic rather than intramuscular or intravenous delivery [31]. Therefore, it seems that transgene delivery via rAAV virus is more successful when using the liver-directed gene route. When intrahepatic therapy was administered to dogs who suffered from hemophilia A, rAAV8 carrying the canine factor VII cDNA showed long-term correction of the phenotype, with no spontaneous bleeding episodes, no toxicity and no development of inhibitory antibodies towards the viral vector or the transgene [32]. Similarly, liver-directed rAAV-FIX therapy to dogs suffering from hemophilia B, significantly increased FIX activity to 4%–10% and remained stable for more than eight years [33]. However, direct injection of the virus to the muscle resulted in unde‐

Among the clinical trials reported in clinicaltrials.gov website, which cites ongoing studies all over the world, the United States is still the leading country conducting clinical trials with rAAV gene therapy. In 2010, 47 studies out of 70 were performed in the United States and in 2015, 44 total studies out of 66 have been sponsored by the same country. Since the first registered trial in 2004, a total of 14 studies have been completed, and three terminated prematurely. Furthermore, there are clinical trials in all phases as well as for traditional, not traditional and even recombinant serotypes (Figure 6). Worldwide, there are a total of more than 130 clinical trials testing rAAV gene therapy for the treatment of diseases (http:// www.abedia.com/wiley/vectors.php). studies out of 66 have been sponsored by the same country. Since the first registered trial in 2004, a total of 14 studies have been completed, and three terminated prematurely. Furthermore, there are clinical trials in all phases as well as for traditional, not traditional and even recombinant serotypes (Figure 6). Worldwide, there are a total of more than 130 clinical trials testing rAAV gene therapy for the treatment of diseases (http://www.abedia.com/wiley/vectors.php).

Therapies for hemophilia B were studied using different routes of rAAV gene therapy administration. Initially, intramuscular delivery of rAAV-

Among the clinical trials reported in clinicaltrials.gov website, which cites

country conducting clinical trials with rAAV gene therapy. In 2010, 47

Figure 6. Statistics showing the clinical trials performed with rAAV **Figure 6.** Statistics showing the clinical trials performed with rAAV gene therapy until 2015 according to clinicaltri‐ als.gov. **A,** Classified according to the clinical phase. **B**, Classified according to the rAAV serotype used in the study.

gene therapy until 2015 according to clinicaltrials.gov. A, Classified

#### 6. rAAV serotypes used in clinical trials according to the clinical phase. B, Classified according to the rAAV **7. rAAV serotypes used in clinical trials**

serotype used in the study.

Traditionally, the most common serotype used in clinical trials is AAV2. In 2010, sixty-two clinical trials were performed with rAAV2 vector; meanwhile, the number was reduced to thirty-six, almost half, in 2015 (89% vs. 54%) (Figure 7A). In addition, in 2010, three studies were performed with rAAV1, which increased to 10 in 2015 (4% vs 16%). Interestingly, more uncommon serotypes are acquiring an interest among the scientific community and the spectrum of serotypes being tested is increasing. Five years ago, five out of seventy clinical trials used serotypes other than rAAV1 and rAAV2. On the contrary, now in 2015, eighteen out of the current forty-nine trials are reported in clinicaltrials.org website (7% vs. 37%). For instance, the number of studies using serotype rAAV8 increased from two to seven in a fiveyear frame (Figure 7B).

The same results were found with the rhesus serotype rAAVrh10; initially, there was one study testing the virus; however, in 2015, six studies have taken place. To note, another rhesus rAAV serotype is being examined: rh74 for duchene muscular dystrophin. Serotypes rAAV5 and

Traditionally, the most common serotype used in clinical trials is AAV2. In

meanwhile, the number was reduced to thirty-six, almost half, in 2015 (89%

with rAAV1, which increased to 10 in 2015 (4% vs 16%). Interestingly, more

Figure 7. Statistics showing the clinical trials performed with rAAV gene therapy in 2010 and in 2015. A, Classified according to the use of traditional serotypes rAAV2 and rAAV1. B, Classified according the

community and the spectrum of serotypes being tested is increasing. Five

vs. 54%) (Figure 7A). In addition, in 2010, three studies were performed

uncommon serotypes are acquiring an interest among the scientific

years ago, five out of seventy clinical trials used serotypes other than rAAV1 and rAAV2. On the contrary, now in 2015, eighteen out of the current forty-nine trials are reported in clinicaltrials.org website (7% vs.

from two to seven in a five-year frame (Figure 7B).

use of no traditional rAAV serotype.

2010, sixty-two clinical trials were performed with rAAV2 vector;

**Figure 7.** Statistics showing the clinical trials performed with rAAV gene therapy in 2010 and in 2015. **A,** Classified according to the use of traditional serotypes rAAV2 and rAAV1. **B**, Classified according the use of no traditional rAAV serotype.

rAAV9 also were introduced into the list of rAAV viruses in clinical trials. Similarly, phase I/ II rAAV10 trial for Sanfilippo type A syndrome started in 2011 and finished in 2014.

As Figure 8A shows, in 2010 a high percentage of the rAAV therapies were in phase I (62%) and a small percentage of the studies (17%) were testing phase I and II on the same trial. In 2015, the number of studies in phase I exclusively was reduced by 20%, compared to studies performed in 2010, and that extra 20% is testing safety and efficacy at the same time (phase I and II, 37%), probably due to the expensive costs of conducting a clinical trial.

Furthermore, the number of studies that were in phase III was reduced, as the therapies started to reach the market. For instance, in October 2012, Glybera became the first rAAV gene therapy to obtain marketing authorization from the European Commission.

Since their discovery in the 1960s as small DNA viruses contaminating cultures of simian and human adenoviruses [2, 34], AAV vectors have been tested in more than a hundred clinical trials. Completed and ongoing trials have consistently confirmed that rAAV vector delivery is safe, well tolerated by humans and efficient in transferring the therapeutic gene. Figure 8B summarizes the spectrum of diseases that have been tested with rAAV gene therapy in 2010

The same results were found with the rhesus serotype rAAVrh10; initially, there was one study testing the virus; however, in 2015, six studies have taken place. To note, another rhesus rAAV serotype is being examined: rh74 for duchene muscular dystrophin. Serotypes rAAV5 and rAAV9 also were introduced into the list of rAAV viruses in clinical trials. Similarly, phase I/II rAAV10 trial for Sanfilippo type A syndrome started in 2011 and

As Figure 8A shows, in 2010 a high percentage of the rAAV therapies were in phase I (62%) and a small percentage of the studies (17%) were testing phase I and II on the same trial. In 2015, the number of studies in phase I exclusively was reduced by 20%, compared to studies performed in 2010, and that extra 20% is testing safety and efficacy at the same time (phase I

finished in 2014.

trial.

Figure 8. Statistics showing the clinical trials performed with rAAV gene therapy in 2010 and in 2015. A, Classified according to clinical phase. B, **Figure 8.** Statistics showing the clinical trials performed with rAAV gene therapy in 2010 and in 2015. **A,** Classified according to clinical phase. **B**, Classified according the treated disease.

and 2015. The statistics show that neurological and ocular diseases are gaining more interest, probably because they both constitute immunological privileged tissues. Figure 9 summarizes the diseases that are being treated with AAV technology, according to the serotype. Classified according the treated disease.

rAAV9 also were introduced into the list of rAAV viruses in clinical trials. Similarly, phase I/

**Figure 7.** Statistics showing the clinical trials performed with rAAV gene therapy in 2010 and in 2015. **A,** Classified according to the use of traditional serotypes rAAV2 and rAAV1. **B**, Classified according the use of no traditional rAAV

Traditionally, the most common serotype used in clinical trials is AAV2. In

meanwhile, the number was reduced to thirty-six, almost half, in 2015 (89%

with rAAV1, which increased to 10 in 2015 (4% vs 16%). Interestingly, more

Figure 7. Statistics showing the clinical trials performed with rAAV gene therapy in 2010 and in 2015. A, Classified according to the use of traditional serotypes rAAV2 and rAAV1. B, Classified according the

community and the spectrum of serotypes being tested is increasing. Five

vs. 54%) (Figure 7A). In addition, in 2010, three studies were performed

uncommon serotypes are acquiring an interest among the scientific

years ago, five out of seventy clinical trials used serotypes other than rAAV1 and rAAV2. On the contrary, now in 2015, eighteen out of the current forty-nine trials are reported in clinicaltrials.org website (7% vs. 37%). For instance, the number of studies using serotype rAAV8 increased

from two to seven in a five-year frame (Figure 7B).

use of no traditional rAAV serotype.

A

130 Gene Therapy - Principles and Challenges

B

serotype.

2010, sixty-two clinical trials were performed with rAAV2 vector;

As Figure 8A shows, in 2010 a high percentage of the rAAV therapies were in phase I (62%) and a small percentage of the studies (17%) were testing phase I and II on the same trial. In 2015, the number of studies in phase I exclusively was reduced by 20%, compared to studies performed in 2010, and that extra 20% is testing safety and efficacy at the same time (phase I

Furthermore, the number of studies that were in phase III was reduced, as the therapies started to reach the market. For instance, in October 2012, Glybera became the first rAAV gene therapy

Since their discovery in the 1960s as small DNA viruses contaminating cultures of simian and human adenoviruses [2, 34], AAV vectors have been tested in more than a hundred clinical trials. Completed and ongoing trials have consistently confirmed that rAAV vector delivery is safe, well tolerated by humans and efficient in transferring the therapeutic gene. Figure 8B summarizes the spectrum of diseases that have been tested with rAAV gene therapy in 2010

II rAAV10 trial for Sanfilippo type A syndrome started in 2011 and finished in 2014.

and II, 37%), probably due to the expensive costs of conducting a clinical trial.

to obtain marketing authorization from the European Commission.

As an ocular AAV therapy, two clinical trials have tested rAAV2 therapy for the correction of Leber congenital amaurosis, an autosomal recessive disease that results in blindness. Specifi‐ cally, patients who participated in these studies received the normal copy of the retinal pigment epithelium-specific 65 (RPE65) gene to correct for the deficient gene. One trial was performed in London and consisted of delivering the gene, the expression of which was driven by an endogenous RPE65 promoter, to adolescent patients [35]. On the other hand, the study performed in Philadelphia delivered the gene in the context of a constitutive promoter, to pediatric and adult patients [36, 37]. Pediatric patients treated in the US resulted in the best improvement in vision, followed by American adults. However, one out of three British patients manifested a visual function improvement. Another trial, sponsored by the University of Pennsylvania, conducted an open-label, dose-escalation phase I study on 15 patients aged between 11 and 30 years. The study examined safety and efficacy. Results showed no toxicity due to the therapy, although some adverse events were observed from the surgery procedure. serotype.

Furthermore, the number of studies that were in phase III was reduced, as the therapies started to reach the market. For instance, in October 2012,

Glybera became the first rAAV gene therapy to obtain marketing

Since their discovery in the 1960s as small DNA viruses contaminating

have consistently confirmed that rAAV vector delivery is safe, well tolerated by humans and efficient in transferring the therapeutic gene. Figure 8B summarizes the spectrum of diseases that have been tested with rAAV gene therapy in 2010 and 2015. The statistics show that neurological and ocular diseases are gaining more interest, probably because they both

constitute immunological privileged tissues. Figure 9 summarizes the

cultures of simian and human adenoviruses [2, 34], AAV vectors have been tested in more than a hundred clinical trials. Completed and ongoing trials

authorization from the European Commission.

Figure <sup>9</sup>. Diseases currently being tested in clinical trials with different rAAV **Figure 9.** Diseases currently being tested in clinical trials with different rAAV serotypes.

Furthermore, visual function was improved in the 15 patients with a variable degree [38]. However, between 9 and 12 months of gene therapy administration, four of the fifteen patients experienced new pseudo-foveas in the retinal regions, for up to six years [39].

The company Sparks, which is sponsoring the studies in the US, is testing the technology developed at the Children's Hospital of Philadelphia in a phase 3 trial and expects to announce their results in 2015. If the results are promising, it could be the next rAAV gene therapy product to be launched in the market.

Among the brain diseases, Parkinson's treatment was attempted with rAAV gene therapy delivering different transgenes. Administration of glutamic acid decarboxylase (GAD) via rAAV2 produced modest efficacy improvements. Patients were injected with rAAV2 coding for GAD65 and GAD67 in the center of the subthalamic nucleus [40]. Six months following the injection, the unified Parkinson's disease rating scale decreased by 8.1 points, compared with a reduction of 4.7 points that the sham operation group evidenced. Six months later, clinical improvements were still being noticed. However, the results were modest and the protocol had some deviations. For instance, patients who showed no benefit on the primary endpoint were eliminated from the statistical analysis, arguing that the injections were off-target [41]. Administration of aromatic L-amino acid decarboxylase (AADC) gene was tested on a phase I trial that consisted of the treatment of 15 patients with moderate disease [42, 43]. The trial, sponsored by Genzyme, observed only a modest efficacy, results that were confirmed by a second study performed in Japan [44]. Similarly, phase I and II trials with the rAAV2-neurturin (CERE-120) vector from Ceregene failed to show statistically significant improvement in the rAAV-treated group compared with the group that received the sham surgery [45, 46]. As a conclusion of all these different trials, the technology is safe and is promising. However, efficacy is modest and does not justify the procedure. Further improvements could be performed, such as modification of the delivery vector, as rAAV1 and rAAV5 are more efficient in transducing the substantia nigra and caudate nucleus than rAAV2. Furthermore, viral dose increase should be considered [41]. On the other hand, long-term improvements were observed during the treatment of Canavan disease [47]. Patients were administered rAAV2 carrying the aspartoaculase gene directly to the brain parenchyma. Five years posttreatment, patients presented slower progression of brain atrophy, fewer frequent seizures and general clinical stabilization. Importantly, no serious adverse events were observed, even when one of the patients was a 3-month-old infant [48].

In a trial testing gene therapy for cardiac disease, patients received different doses (low, medium or high) of the sarcoplasmic reticulum Ca2+-ATPase (SERCA2a) gene via rAAV1 [49]. Six months following the percutaneous intracoronary infusion of the virus, several clinical parameters, such as walk test, peak maximum oxygen consumption, left ventricular endsystolic volume, cardiovascular events and time to clinical events were stabilized or even improved. Currently, a phase 2b trial is ongoing, which would test a larger patient population (NCT01643330).

The first clinical trial for rAAV gene therapy that reached the market was the product Glybera® (alipogene tiparvovec), an rAAV1 vector delivering a lipoprotein lipase variant (LPLS447X) for the treatment of lipoprotein lipase deficiency (LPLD). Lipoprotein lipase is a secreted enzyme produced by the skeletal muscle and adipose tissue. Its function involves the metab‐ olism of triglycerides, chylomicrons and very low-density lipoproteins. Three clinical trials showed that Glybera is safe and efficient for the treatment of LPLD. In the first trial, two doses of rAAV1\_LPLS447X were studied: low and high [50]. Nevertheless, none of the doses resulted in a permanent decrease in triglyceride levels. There was only a transient reduction, possibly due to the development of an immune response. The second clinical trial received the therapy in combination with an immunosuppressive regimen [51]. Similar to the first clinical trial, the effects of the therapy were only transient in the beginning. However, improvements were observed after two years posttreatment, such as tolerance to certain foods, changes in the blood lipid content and a decreased frequency of pancreatitis. Due to a discrepancy in the clinical outcomes and plasma triglycerides levels, a third trial was set with predetermined parameters to measure, as incidence frequency of abdominal pain, pancreatitis and chylomicron plasma clearance [52]. Five newly treated patients evidenced a reduction of the parameters and an improved quality of life for two years following administration.

These results, combined with the ones obtained from the reanalysis of 22 of the 27 previously treated patients, confirmed the therapeutic benefits of therapy and granted its approval to the market by the European Commission in November 2012.
