Progress and problems when considering gene therapy for GSD-II

Introduction

Glycogen storage disease type II (GSD-II), also known as Pompe disease, or acid maltase deficiency (AMD), is an autosomal recessive genetic disorder that encompasses a range of clinical phenotypes, but myopathy is common to all. This “variable expressivity” manifests primarily as variances in age of onset of disease symptoms, as well which organs are pathologically involved. The most severe form of GSD-II is the infantile-onset form, and was originally described by Dr. Pompe. These severely affected infants may appear normal at birth, but soon develop generalized muscle weakness and cardiac myopathy manifesting initially as a hypertrophic cardiomyopathy. Shortly thereafter death resulting from cardiorespiratory failure typically occurs before the age of 2 years. The juvenile (onset as early as at age 1 year) or adult onset (onset between the second or later decades) forms of GSD-II present as slowly progressive limb muscle myopathies, and lack the cardiac involvement noted in the infantile patients. The clinical picture of these later onset forms of GSD-II are dominated by a slowly progressive respiratory muscle and proximal limb muscle weakness, with truncal involvement and greater involvement of the lower, rather than the upper limbs. Although these forms of GSD-II are not lethal in the neonatal period, juvenile and adult GSD-II patients suffer from significant morbidity and mortality, the latter primarily due to the complications of respiratory insufficiency.

GSD-II is caused by the inheritance of mutant alleles that either result in the complete lack of expression of acid alpha-glucosidase (GAA) protein; this is the protein which breaks down intra-lysosomal glycogen. In general, the severity of the clinical phenotypes can be correlated with the residual GAA enzyme activity levels measured in a respective patient’s tissues. For example, infantile GSD-II patients typically have less than 1% of normal GAA activity levels in their muscles, while juvenile or adult onset forms of GSD-II may have 2-40% of normal GAA tissue activity levels. Infantile patients may have nonsense mutations that prevent any GAA protein from being expressed, or missense mutations that allow for production of an enzymatically “dead” GAA protein. Juvenile or adult onset GSD-II patients, have less severe mutations (missense or splice-site mutations) that cause expression of a less than nominal GAA protein, or decreased levels of a normal GAA protein. These facts alone demonstrate that very low levels of GAA activity allow for preservation of normal cardiac function in juvenile and adult GSD-II patients. However, in juvenile and adult onset patients, GAA activity levels generally cannot be positively correlated with rate of progression and/or disease severity of respiratory or limb muscle involvement, suggesting that other genes and/or environmental factors likely significantly impact on disease severity in juvenile or adult onset GSD-II patients. The true incidence of GSD-II (in all its presenting forms) is not accurately documented, but estimates are in the range of 1 in 40,000 to 1 in 100,000 live births. Due to the relatively rare occurrence of a GSD-II diagnosis, GSD-II has been designated an orphan disease.

Prior Attempts at Therapy for GSD-II

The discovery of cell-surface receptors that can mediate the delivery of lysosomal enzymes into target tissues has given promise to the use of enzyme replacement therapy (ERT) for treatment of GSD-II (1–3). Our group reported the first US study to demonstrate efficacy of recombinant human GAA (rhGAA) enzyme infusions in infantile GSD-II patients, with both cardiac and skeletal muscles responding (4). While rhGAA could offer a promising therapy for GSD-II, several problems with ERT approaches have also been noted. These include difficulties in large scale production, the high cost of the recombinant enzyme, the need for repeated and life-long infusions, as well the possibility that current dosing regimens may not be adequate to treat all muscle cells in an affected individual. These realities have prompted our research interest into alternative therapeutic options for GSD-II patients. Past and ongoing developments within our laboratories show the promise of virus mediated gene transfer of the hGAA gene for potential therapeutic use in all GSD-II patients.

Gene Therapy for GSD-II

Clearly, any gene transfer “vector” will have advantages and limitations depending on the specific requirements of the gene therapy application. Virus based vectors are generated by recombinant DNA technologies, in which deleterious virus genes are removed, and replaced with desirable gene products, such as the hGAA gene (5). Use of appropriate packaging cell lines allows high level growth of the recombinant vectors in tissue culture, and the vectors are purified to high concentration for use in animal models. The two predominantly used gene transfer vectors for GSD-II treatment are Adeno-associated viruses (AAV) and Adenoviruses (Ads). These viral vectors have been confirmed to be able to “transduce” the human GAA (hGAA) gene to target cells in tissue culture systems, as well several animal models of GSD-II.

Adenovirus (Ad) based gene therapy for GSD-II: Ad based vectors are one of the best characterized gene transfer vectors; their numerous benefits have made them widely useful in basic biology studies, cell and gene therapy applications, as well for vaccine development (6–9). Using an in vitro model, it was demonstrated that an Ad vector could mediated transfer of a GAA gene into cultured fibroblasts and myotubes from a GSD-II patient, and this resulted in: 1) de novo synthesis of GAA enzyme, 2) clearance of lysosomal glycogen, and 3) secretion of 110 kDA GAA being taken up by recipient cells (10). Subsequently, in vivo studies showed that there was insufficient secretion of muscle derived hGAA to allow cross-correction of non-transduced muscle cells (11–13). For example, Pauly and colleagues demonstrated that intramuscular injection into cardiac or skeletal muscles of normal neonatal rats of a first generation Ad vectors expressing hGAA resulted in high levels of GAA expression only within the injected muscles, but systemic correction of non-injected muscles was not achieved (14).

Obviously, any therapeutic strategy for GSD-II should allow for treatment of all muscle cells in an affected individual. This is a significant hurdle to surpass, as up to 40% of one’s mass may be muscle tissue. Due to this significant limitation, we set out to create a system that could achieve this goal. Our group first investigated this problem by utilizing direct intravenous injection of hGAA encoding Ad vectors, capitalizing on the knowledge that the liver is primarily targeted by this mode of vector administration (15, 16). GAA-KO mice injected with an advanced generation Ad vector began expressing, and more importantly, secreting high levels of precursor hGAA from their transduced liver hepatocytes. Since this form of hGAA is likely correctly processed by the mammalian Golgi apparatus to contain the “motifs”: (i.e.: mannose-6-phosphate residues) required for receptor mediated uptake and lysosomal targeting, liver derived hGAA was taken up by all affected muscle cells in the Ad treated GAA-KO animals. This resulted in both cardiac and skeletal muscle glycogen clearance within 12 days of the gene therapy treatment, and like ERT, was most effective in cardiac and diaphragm muscles (17). This report was the first to demonstrate the complete systemic correction of a form of muscular dystrophy by a simple intravenous injection technique. This observation has been subsequently repeated by us and others using Ad or AAV based approaches, as well in alternative species, such as the AMD quail model of GSD-II (18–20).

Long term Ad mediated hGAA gene therapy studies showed that hepatically derived hGAA enzyme persisted in the heart and diaphragm for at least 6 months post injection. However, we also noted that plasma levels of hGAA diminished over time in the vector treated GAA-KO mice (21). We have subsequently confirmed in numerous publications that the loss of plasma hGAA following intravenous injection of an Ad-hGAA vector was not due to loss of the vector, but rather was due to the onset of anti-hGAA antibodies (11, 22–24). These anti-hGAA antibodies prevent efficient, high level skeletal or cardiac muscle cell uptake of hepatically expressed hGAA, and limited long term efficacy of the approach, a limitation also noted after ERT treatments of GSD-II patients (4, 25). In our most recent studies, use of an optimized Adenovirus vector expressing hGAA via a tissue specific promoter in adult, hGAA tolerant GAA-KO mice permitted glycogen correction and muscle strength to be preserved in hGAA tolerant, GAA-KO mice for greater than one year. This result was primarily achieved by avoiding the elicitation of anti-hGAA antibody production in the treated animals (26).

To begin to investigate the potential for disease reversibility in older GSD-II mice, we studied their responsiveness to hGAA gene therapy. This became especially relevent, since ERT in older mice had been shown to have significant limitations (27). Upon intravenous Ad-hGAA vector injection, we found that extremely high amounts of hepatically secreted hGAA could be produced, and subsequently taken up by multiple muscle tissues in old (12-17 months) GAA-KO mice. As a result, all muscle groups in the hGAA gene treated mice showed significant glycogen reductions, relative to that of age-matched, but mock-injected old GAA-KO mice. Our data also showed that the uptake and the subsequent intra-cellular processing of virally expressed hGAA was not impaired in older muscles. Thus the extremely high levels of hGAA expression afforded by use of an Ad mediated gene therapy approach overcame many of the issues noted with ERT approaches when tested in similar, animal models.

High dose intravenous injection of Ad vectors is not without untoward side effects. For example, at extremely high doses, Ad injection ultimately results in generation of high-titer, neutralizing anti-Ad capsid antibodies, that prohibit re-infection of many tissues with same serotype Ad vectors (28, 29). Furthermore, after high dose injection, the Ad capsid proteins themselves appear to immediately elicit “innate” immune responses, such as increased plasma cytokine and chemokine levels and activation of the complement system (30–32). Many of these same responses have been noted after [E1-]Ad injections into non-human primates, and humans (33).

Initial studies of AAV vectors demonstrated that the efficacy of AAV-hGAA vectors was hampered by choice of AAV serotype and promoter used, low-level hGAA expression, and production issues that still limit large-scale AAV production (34–36). Furthermore, it was noted that upon intramuscular injection of certain AAV serotypes, the virus could be found in uninjected sites, suggesting that the virus could cross normal tissue barriers (18, 37, 38). Recently, intravenous administration of newer serotypes of AAV, such as AAV serotype 8 expressing hGAA from a liver-specific promoter, resulted in high level GAA enzyme production, glycogen reduction in some muscle groups, and minimization of the anti-hGAA humoral immune responses normally noted in GAA-knockout (KO) mice treated with vectors expressing hGAA from non-viral promoter elements (39). Also AAV-9 has shown an improved ability to infect and allow for expression of hGAA genes from cardiac tissues in vivo (40).

AAV vectors can feasibly be used for GSD-II treatment, but these vectors face similar issues as those noted with Ad based vectors, issues that have become more detectable as the titer (and thus therapeutic potential) of AAV preparations have increased. These include cytokine responses, and elicitation of anti-AAV specific antibody responses (41). Some further considerations include the oncogenic potential (due to integration) and possible germline transmission of AAV vectors, and a limited genetic carrying capacity (< 5 kb) (42–46). Finally, high dose intravenous administration of alternative serotype AAV based vectors allows for dissemination of the vector beyond endothelial barriers into not only muscle tissues, but also gonadal sites (43), (47). Whether this property is a benefit or limitation (i.e. what is the mechanism for this capability?) awaits further research.

In conclusion, gene therapy strategies for GSD-II have demonstrated a number of potential benefits when tested in several animals models of GSD-II. These benefits include, 1) capability for long term hGAA gene transfer and expression after a single treatment, 2) the first demonstration that systemic correction of a muscular myopathy could be achieved after a simple intravenous administration, 3) that the abnormal glycogen storage in multiple muscle groups can be corrected by using the liver as a hGAA secretory depot organ, and 4) that long term muscle correction with physiologic improvement, as well correction of glycogen storage in long standing diseased muscles have and can all be achieved via gene therapy approaches. In many instances these results rival, or exceed the capabilities of ERT approaches tested in similar models. Furthermore, gene therapy research in GSD-II has shed light on the complexities of the host immune response when exposed to potentially foreign proteins such as hGAA, although aspects of gene therapy (such as using tissue specific promoters, especially in the context of an hGAA tolerant animal) suggest that these limitations can also be overcome with gene therapy approaches. However, the numerous acute and chronic risks currently associated with gene therapy vectors may limit its use to only the most severely affected GSD-II patients, (i.e.: those which don’t respond to ERT). Future research in gene therapy for GSD-II should thus focus on understanding and overcoming the toxicities associated with in vivo gene transfer, as well as potentially utilizing combined ERT/gene therapy approaches to synergistically improve the efficacy and/or decrease the toxicity of either form of therapy.