Gene Therapy for Rare Diseases

Barry Byrne, pediatric cardiologist from the University of Florida’s Powell Gene Therapy Center, opened the symposium by describing his ongoing efforts to treat infantile-onset Pompe disease. This inherited disorder is caused by mutations in a gene that encodes the enzyme alpha-glucosidase, which normally degrades glycogen in lysosomes. In the absence of this enzyme, muscle tissue is overrun by glycogen-laden lysosomes that interfere with cellular ion channels, altering muscle structure and function and leading to muscle weakness, enlarged hearts, and early mortality. At first, Byrne and his colleagues treated the problem by directly replacing the missing enzyme—with initially encouraging results; all treated infants survived with normal heart function past 1.5 years of age, as compared to the one-in-five survival rate among untreated infants. Extending the lifetime of the treated patients, however, allowed the investigators to discover another related, but previously unnoticed, deficit in diaphragm function. Though children were surviving longer, they had difficulty breathing and required assisted ventilation. Byrne and colleagues determined this was due to decreased innervation of the muscles by motor neurons. To treat diaphragm dysfunction, Byrne turned towards a novel method that his group had been developing in parallel with others since the 1990s—the use of adeno-associated viruses (AAVs) to deliver a gene to the body. Byrne and his colleagues reasoned this method—using AAVs to replace the alpha-glucosidase gene instead of just the protein—would offer the potential for longer term, more stable expression of the necessary protein. As a first step towards treating diaphragm dysfunction in Pompe patients, the group injected AAV vectors into knockout mice designed to exhibit progressive accumulation of glycogen. In these mice, vectors introduced to the muscle successfully delivered the missing gene into the cell bodies of motor neurons, a process illustrated when the neurons stained positive for gene expression. In addition, the quality of neurons’ neurophysiological signals were restored to wild-type levels. To measure the effect of the treatment in clinical trials, Byrne and his team measured airway flow in patients who required ventilator assistance. Following an injection of an AAV vector carrying the alpha glucosidase gene directly into the diaphragm muscle, unassisted breathing in Pompe patients improved, even long after treatment. Patients with the most severe dysfunction—who previously returned to deficient airway states only minutes after being removed from a ventilator—a year later could breathe independently for several hours.

Brian Kaspar of AveXis Inc. spoke of his company’s progress toward developing gene therapy for spinal muscular atrophy (SMA), a monogenetic disorder caused by deletions in the survival motor neuron (SMN) gene and subsequent lower levels of SMN protein, which are crucial for normal motor function. To replace the gene, Kaspar and colleagues cloned the SMN gene and reintroduced it into the patients’ cells using AAV serotype 9, which can be delivered systemically while still targeting motor neurons. AAV9 treatment was also remarkably efficient: in mice, over 90% of motor neurons were targeted by just a single injection. Furthermore, preclinical trials found that even a low dose of the SMN gene could double the lifespan of animals who would otherwise normally die within 15 days. Recently, the company completed phase I and II clinical trials to deliver the gene to 15 SMA patients with SMN gene deletion, reporting that, as of the study’s 13-month endpoint, all patients were event-free. To measure improvements in motor function, researchers also developed a battery of 16 tests to assess a variety of motor functions, such as the ability to roll over or bring hand to mouth. Patients showed a marked improvement in reaching motor milestones after treatment.

In yet another demonstration of how AAV viruses can be used to replace defective genes, Kevin Flanigan, director of the Nationwide Children’s Hospital, described his team’s preclinical and clinical studies on the inherited metabolic disorder Sanfilippo syndrome. Children suffering from Sanfilippo lack one of four enzymes due to defects in the corresponding genes, and the type (A through D) of the syndrome depends on which of the four is affected. The most common forms, types A and B, involve missing or deficient quantities of the enzymes heparan-N-sulfatase and alpha-N-acetylglucosaminidase, respectively. This leads to an inability to properly break down the sugar molecule glycosaminoglycan during normal metabolism, resulting in patients with enlarged livers and spleens. While normal at birth, these children begin to exhibit speech delay and behavioral and cognitive problems in early childhood; death usually follows by the late teen years. Flanigan and colleagues used a systemic delivery of AAV9 in mice to show that they could transduce cells, primarily in the liver and spleen, with the gene responsible for the most common type of Sanfilippo syndrome. In clinical trials, they noted that patients who received treatment had a significant decrease in liver and spleen volume by 30 -180 days. In addition, patients exhibited improved nonverbal IQ scores, as compared to individuals described in natural history studies that have tracked the characteristics and time course of the disease.

Jakub Tolar, director of the Stem Cell Institute at the University of Minnesota, illustrated how the CRISPR/Cas9 gene editing method might come into play in this arena. CRISPR, as distinct from gene therapy, involves altering the sequence of defective DNA to restore its normal function. Currently, Tolar is using this method to treat a condition called Fanconi anemia, a disease which typically hits children around 8 years of age and is marked by a failure of hematopoiesis in the bone marrow, caused by defects in a cluster of proteins necessary for DNA repair. With the CRISPR/Cas9 method, Tolar and his team targeted the corresponding sequence of mutated genes, replaced them with a sequence of properly encoded genes, and thus restored the function of the protein complex. He is now working on amplifying the effect of gene editing by reprogramming patients’ edited skin cells into embryonic stem cells, with the goal of ultimately delivering these reprogrammed functional cells back to the same individual, and thereby rendering viral delivery unnecessary.

While numerous labs are finding success with virally-mediated gene therapy, researchers have also been exploring delivery method refinements and other particularly promising new directions for expanding treatment options.

Jude Samulski of Bamboo Therapeutics stressed that while much of the focus currently lies on the gene carried by the virus—as the therapeutic agent to be delivered—the capsid—the virus’s protein shell—offers another means for boosting effectiveness. In order to tease out some of the possibilities, he and his colleagues conducted numerous studies to understand variability in capsid structure and function. Viral serotypes are distinguished by amino-acid motifs present on their surfaces, which interact with cell receptors. This mechanism is what determines whether a virus binds to, say, a liver cell or a neuronal cell. For example, AAV1 is best suited for intramuscular targets, AAV2 to those in the eye, and AAV8 to liver cells. Using this knowledge, Samulski and colleagues have been able to advance numerous trials using virally-mediated gene therapy. In the treatment of muscular dystrophy, his group wanted to find the serotype that would best improve skeletal muscle transduction. While most data from clinical trials had been completed using an AAV2, AAV1 was more suited to expression in muscle. They therefore identified five amino acids from the capsid sequence of AAV1 they predicted would have the biggest impact in targeting muscle cells, and combined them with AAV2. The resulting chimeric virus, AAV 2.5, had significantly more efficient muscle transduction. This resulted in the first example of a chimeric AAV vector being tested in a clinical setting.As an outcome, this novel approach has paved the way for the numerous new laboratory-derived AAV vectors to be tested in future clinical settings.

In addition to viral modification, the route of administration is also an important consideration. Steven Gray, assistant professor at University of North Carolina at Chapel Hill, described two methods compared by his group: intravenous versus intrathecal delivery into the cerebrospinal fluid. Comparing the biodistribution of a gene carried by AAV9 through each route, Gray observed that intrathecal injections tend to lead to higher concentrations in the spinal cord and organs such as the liver and kidney, but lower ones in the brain, while intravenous injections resulted in expression throughout the periphery, with lower levels in the nervous system and a high concentration in the liver.

Because many rare genetic diseases manifest in infants and young children, the earliest possible intervention is optimal. At the same time, however, assessing improvements in function after gene therapy is difficult in this population, as even in normal development, brain function is constantly evolving.

Maria Escolar, neurodevelopmental pediatrician and clinical researcher from the University of Pittsburgh, spoke about her efforts to examine the natural history of lysosomal storage disorders in children, a project designed to drive selection of appropriate biomarkers for comparing treatment recipients to affected and healthy subjects. For instance, the brain reaches 80-90% of the adult brain volume after just two years of age—pediatricians typically measure head growth during this stage because it correlates with brain growth. But this metric does not capture the complex changes in connectivity that occur during development. Moreover, most pediatric tools that test for cognitive ability are designed to capture one-time assessments, rather than longitudinal ones. Thus, if these metric snapshots appear to indicate that a child with lysosomal storage disorders is below normal when in fact the patient is learning but at a slower speed than normal, clinicians might assume the child is not improving overall. To help rectify this, her group created a practical guide laying out more nuanced ways to understand development of these children. They found that a more effective way to assess development is to do repeated measures  over time and to factor in numerous, diverse measurements of developmental equivalence that took into account different areas of development expressive language, gross and fine motor function, and cognitive ability, among others  to help identify impact on several functional areas. Diffusion tensor imaging is one such tool that Escolar and her group used in natural history studies of the diseases. Imaging the corticospinal tract, a bundle of nerve fibers that relays action potentials from the brain to the spinal cord, can reveal the shape and integrity of the tract in infants within their first weeks. In infants with Krabbe’s disease, which results in severe motor dysfunction within three months, the severity of symptoms and effectiveness of treatment could be predicted by the diminished integrity of the corticospinal tract. Through quantitative and standardized imaging, doctors could greatly enhance the likelihood of successful treatment by interceding long before symptoms become visible through other measures.

In the final presentation Katherine High of Spark Therapeutics, described the process of developing gene therapies for genetic disorders with no previously established metrics or a history of treatment. Inherited blindness, caused by mutations in the retinal pigment epithelia 65 (rpe65) gene is a disease that—unlike other rare diseases—had neither existing therapeutics nor any natural history studies. The rpe65 gene normally codes for an enzyme required in the regeneration of cis-retinal, a molecule photoreceptors rely upon for normal vision function. In preclinical trials on canine models of rpe65-mediated blindness, injecting AAV2 carrying the rpe65 gene into the sub-retinal space restored vision. During clinical trials however, many questions arose as to what the appropriate controls were, how to estimate the size of a phase III study, and what to do about the lack of natural history studies examining the normal disease progression. High and her group thus collected a wide variety of outcome measures, such as visual acuity and visual field, in addition to developing a new metric: a multi-luminance mobility test, to assess functional vision. This test required patients to navigate a maze of physical barriers under varying levels of illumination, ranging from a mere nightlight to a brightly lit office. Researchers then compared the dimmest level at which individuals passed before treatment and a year after. In addition, they conducted the battery of tests in groups with normal vision, with other visual impairments, and in those who had not yet received treatment. Tested a year after gene therapy, patients receiving treatment passed the mobility test at a significantly lower light intensity.