Less than a year after reporting preclinical proof of concept for a systemically delivered mucopolysaccharidosis IIIA gene therapy, researchers at the Autonomous University of Barcelona have shown that intracerebral delivery can substantially improve efficacy in the CNS and treat symptoms in peripheral organs.1 Esteve S.A., who partially funded the work, is now raising funds to test the approach in the clinic.

Mucopolysaccharidosis IIIA (MPS IIIA)-also called Sanfilippo type A syndrome-is caused by a mutation in N-sulfoglucosamine sulfohydrolase (SGSH) that impairs the ability of SGSH to degrade heparan sulfate glycosaminoglycans (HSGAGs).

The resulting accumulation of HSGAGs in the brain causes a range of motor, behavioral and other neurological symptoms, whereas accumulation in the periphery can lead to enlargement of the liver and spleen. There are no therapies approved to treat the disease, and most patients with MPS IIIA die in their mid-teens.

Therapies in development to treat MPS IIIA include enzyme replacement therapy with SGSH, administered either systemically by i.v. injection or intrathecally by a surgically implanted pump.

Both delivery methods have drawbacks. Systemically administered enzyme cannot penetrate the blood brain barrier (BBB) to treat CNS symptoms. Permanent intrathecal implants can malfunction and induce infections at the implantation site.

Autonomous University of Barcelona (UAB) and Esteve have been exploring alternative approaches. The group focused on adeno-associated virus serotype 9 (AAV9)-based SGSH gene therapy after about a decade of research from multiple groups showing that AAV9 isolated from human tissues2 could cross the BBB in both directions.3,4

In 2012, the UAB team reported that systemic administration of an AAV9 vector encoding mouse Sgsh-induced expression of the enzyme in the brain and normalized HSGAG levels in peripheral organs.5

The problem was that the maximum dose of the therapy resulted in Sgsh activity in the brain that was only 10%-12% of normal levels and larger doses would have induced significant liver toxicity, team leader Fátima Bosch told SciBX.

For the new study, Bosch's team postulated that smaller doses of the therapy injected directly into cerebrospinal fluid (CSF) would have comparable or greater CNS efficacy than systemic administration while safely treating disease symptoms in peripheral organs.

First, the team administered the AAV9 vector encoding murine and canine Sgsh to mouse models for MPS IIIA and normal dogs by injecting it into the cisterna magna, a space within the cerebellum that is filled with CSF (intracisternal injection). Compared with empty vector, the therapy increased Sgsh activity in the brain to about 40% of normal levels and decreased HSGAG levels in the spleen, kidneys and other organs. In mice, the therapy decreased behavioral deficits and increased locomotor function and survival compared with empty vector or no treatment.

Importantly, the team achieved significant CNS efficacy at doses up to 20-fold lower than the i.v. doses used in its 2012 study.

In normal dogs, intracisternal and intracerebroventricular injections of the AAV9 vector encoding GFP resulted in comparable distributions of the protein throughout the brain and peripheral organs. This finding suggested administering the therapy via clinically established intracerebroventricular procedures would be just as efficacious as intracisternal delivery, which is not commonly used in clinical practice, the team wrote in its report in The Journal of Clinical Investigation.

"This is the first demonstration of the whole-body efficacy for a gene therapy to treat a lysosomal storage disorder," Bosch said.

Bosch is director of the Center of Animal Biotechnology and Gene Therapy at UAB. Her team included researchers from the St. John of God Hospital and The Children's Hospital of Philadelphia. Esteve partially funded the study but did not directly participate in it.

Surgical strike

Michaël Hocquemiller, senior scientific affairs program manager at Lysogene, said, "The main theoretical advantage of intracerebroventricular delivery is that one could treat the disease in the periphery in addition to the CNS."

He cautioned that intracerebroventricular delivery of gene therapy is still at an exploratory stage and that it is unclear how it compares with the intraparenchymal delivery method that Lysogene's SAF-301 gene therapy uses.

In May, Lysogene completed a Phase I/II study of SAF-301, an intraparenchymally administered AAV10 vector encoding SGSH and sulfatase modifying factor 1 (SUMF1), to treat MPS IIIA.

Karen Aiach, cofounder, chairman and CEO of Lysogene, said that Marc Tardieu, principal investigator of that trial, plans to present safety and efficacy data from the trial at the European Society of Gene and Cell Therapy conference this October.

Tardieu is professor of pediatrics and head of the pediatric neurology unit at the University Hospital of Bicêtre.

The company will also report the trial results in a forthcoming paper, Aiach said.

Intraparenchymal delivery involves drilling six burr holes in the skull and injecting a therapy directly into the brain matter at two different depths in each hole. Intracerebroventricular delivery involves drilling one burr hole and injecting a therapy into CSF contained in brain ventricles.

Olivier Danos, cofounder and senior advisor at Lysogene, said that the new study "showed that significant amounts of vector had to be used for injection in the cerebrospinal fluid."

The equivalent dose of the AAV9-based therapy in children would be "more than 100 times greater than the one we used in our clinical trial of SAF-301," Danos said. Thus, future clinical application of the team's therapy would be limited by the ability of GMP manufacturers to produce sufficient vector, he said.

Danos added that the safety of the large doses of vector injected in the CSF would have to be established.

Danos also is SVP of molecular medicine, synthetic biology and gene regulation at Kadmon Corp. LLC's Kadmon Pharmaceuticals LLC unit, associate professor at the University College London's Cancer Institute and research director at the Centre National de la Recherche Scientifique (CNRS).

However, Mark Mayhew, director of strategic development at Esteve, said the doses used in the JCI study would not present the company with manufacturing difficulties.

Moreover, "the ratio of our proposed dose to the volume of CNS target tissue, which includes the whole brain, cerebellum, brain stem and spinal cord, is not higher than those previously shown to be safe in clinical trials" of Leber's congenital amaurosis (LCA) and Parkinson's disease (PD), he said.

Indeed, the vector injected into the CSF distributes evenly across the CNS and periphery, resulting in vector levels that are much lower than those found near intraparenchymal injection sites, he said.

Bosch pointed out that-as reported in the JCI paper-her team observed no signs of CNS inflammation or other safety problems in a seven-day follow-on study in normal dogs that received intracisternal injections of AAV9 vector encoding canine SGSH.

Additionally, the intracerebroventricular injection is already used to treat hydrocephalus (abnormal accumulation of CSF in brain ventricles) and deliver antibiotics, pain therapies and cancer drugs in 40,000-50,000 pediatric patients in the U.S. each year, giving it a safety track record that intraparenchymal delivery does not yet have, said Eduard Valenti, Esteve's director of regulatory affairs and pharmaceutical quality.

"Our approach uses a standard procedure that most neurosurgeons have the skills to perform," Bosch added.

Intraparenchymal delivery requires special expertise and a special injection device that most pediatric surgeons lack, she said.

"Intracerebroventricular delivery is less invasive than intraparenchymal delivery and gets the therapeutic vector into the ventricle through a part of the brain that does not house important functions, so the risk of damage is minimal," said Virginia Haurigot, a research associate in Bosch's group at UAB and first author on the JCI paper. "Also, direct delivery into CSF allows the therapy to reach parts of the CNS that are hard to reach with intraparenchymal injection."

Both Haurigot and Mayhew said the efficacy of intraparenchymal gene therapy was not yet established.

Haurigot said that published studies in animals have shown that intraparenchymally delivered vectors have limited diffusion in the CNS. "Therefore you cannot reach very deep structures of the brain such as the cerebellum and brain stem, and transfection of the brain stem appears to correlate with efficacy" for at least some CNS diseases, such as Canavan disease,6 she said.

Mayhew added, "It's also not clear whether intraparenchymal therapies can cross the blood brain barrier and get into the periphery."

According to Hocquemiller, at least two studies have shown that i.v. AAV10 crosses the BBB as efficiently as AAV9.7,8 Although he had not seen any studies showing that an intraparenchymally delivered vector could cross the BBB, "if that did happen, the amount would probably not be sufficient to treat symptoms in peripheral organs," he said.

Different development vectors

Hocquemiller and Danos both said that the results of the JCI paper raise another obstacle for the AAV9 vector that AAV10 does not have.

"The study shows that the presence of pre-existing serum antibodies against the AAV9 vector is likely to block the vector's transduction of peripheral organs," thereby reducing its efficacy in those organs, Hocquemiller said.

In the study, the team injected an AAV9 vector encoding GFP into dogs that were preimmunized against AAV9. Whereas the low levels of anti-AAV9 antibodies in CSF did not significantly affect the distribution of GFP in the brain, the high levels of serum antibodies did reduce the ability of the vector to reach peripheral organs and express GFP.

"Up to 80% of humans have serum antibodies against AAV9, AAV2 and/or AAV5, which may limit the use of these as vectors for gene therapy," Danos said. "Lysogene's AAV10 vector, which is derived from a rhesus money, may be more applicable to gene therapy" because antibodies against it have not been found in humans.

Haurigot acknowledged that "some efficacy was lost in the periphery due to serum anti-AAV9 antibodies" but noted that there was no illness or adverse event associated with them.

But because MPS IIIA is primarily a neurodegenerative disease, "the main target of the therapy is the brain, and we still achieved efficacy in the CNS even when the dogs had pre-existing immunity to AAV9," she said.

Haurigot added that the likelihood of developing anti-AAV9 antibodies increases with age, but young children comprise the target population for MPS IIIA therapies. Thus, "a large proportion of patients may benefit from the peripheral efficacy of our therapy because they would not have been previously exposed to AAV9," she said.

Going forward, Esteve is seeking investors or corporate partners to help fund a Phase I/II trial of intracerebroventricular AAV9-based SGSH gene therapy to treat MPS IIIA. The start date of the trial would depend on when the company secures the necessary funding, Mayhew said.

Mayhew added that ReGenX Biosciences LLC owns IP around many AAV vectors, including AAV9. Thus, "the entity that eventually commercializes our gene therapy would have to come to some arrangement with ReGenX," he said.

Meanwhile, Bosch's team is testing AAV9-based gene therapy in animal models of MPS IIIB (Sanfilippo type B syndrome), a disease that is caused by a mutation in N-acetylglucosaminidase-a (NAGLU), and other types of MPS.

UAB and Esteve hold a patent for the AAV9-based gene therapy to treat MPS IIIA, and the IP is exclusively licensed to Esteve.

Haas, M.J. SciBX 4(28); doi:10.1038/scibx.2013.709
Published online July 25, 2013


1.   Haurigot, V. et al. J. Clin. Invest.; published online July 1, 2013; doi:10.1172/JCI66778
Contact: Fátima Bosch, Autonomous University of Barcelona, Bellaterra, Spain
e-mail: fatima.bosch@uab.es

2.   Gao, G. et al. J. Virol. 78, 6381-6388 (2004)

3.   Foust, K.D. et al. Nat. Biotechnol. 27, 59-65 (2009)

4.   Duque, S. et al. Mol. Ther. 17, 1187-1196 (2009)

5.   Ruzo, A. et al. Hum. Gene Ther. 23, 1237-1246 (2012)

6.   Janson, C. et al. Hum. Gene Ther. 13, 1391-1412 (2002)

7.   Hu, C. et al. J. Gene Med. 12, 766-778 (2010)

8.   Zhang, H. et al. Mol. Ther. 19, 1440-1448 (2011)


      Autonomous University of Barcelona, Bellaterra, Spain

      Centre National de la Recherche Scientifique, Paris, France

      The Children's Hospital of Philadelphia, Philadelphia, Pa.

      Esteve S.A., Barcelona, Spain

      European Society of Gene and Cell Therapy, Heidelberg, Germany

      Kadmon Corp. LLC, New York, N.Y.

      Lysogene, Paris, France

      ReGenX Biosciences LLC, Washington, D.C.

      St. John of God Hospital, Barcelona, Spain

      University College London, London, U.K.

      University Hospital of Bicêtre, Le Kremlin-Bicêtre, France