Brain-directed gene therapy for lysosomal storage disease: Going well beyond the blood– brain barrier

The lysosomal storage diseases (LSDs) are a heterogeneous group of disorders that affect 1/7,000 live-born infants, the majority of which develop central nervous system (CNS) disease. Brooks et al. (1) report exciting results from Davidson's group with brain-directed gene therapy for murine mucopolysaccharidosis (MPS) VII that are likely to have general implications for the treatment of CNS disease in LSD. Each LSD results from a deficiency of a single lysosomal enzyme important for degrading macromolecules that must be turned over in lysosomes. More than 40 LSDs have been described (2). Over the past two decades, dramatic progress has been made in understanding the biogenesis, structure, and function of lysosomes and the processes by which newly synthesized acid hydrolases are assembled, processed, and transported to lysosomes.

Understanding the receptors that target enzymes to lysosomes led to the development of successful enzyme replacement therapy.

Understanding the receptors that target enzymes to lysosomes, some of which are expressed on the cell surface, led to the development of successful enzyme replacement therapy for one of the LSDs, Gaucher Disease, a disorder of sphingolipid degradation (3). Gaucher Disease results from deficiency of glucocerebrosidase (β-glucosidase), the enzyme involved in the last step of sphingolipid degradation. Storage of glucocerebroside in macrophages produces tremendous enlargement of spleen and liver, disabling bone involvement and occasional pulmonary incapacity. The strategy for treatment involved purification of placental enzyme and later recombinant enzyme from Chinese hamster ovary cell secretions and modification of the native enzyme to expose mannose residues on oligosaccharides. This strategy targets the infused enzyme to the mannose receptors of fixed-tissue macrophages, precisely the cells affected by the storage; receptor-mediated endocytosis delivers enzyme to the lysosomes where the substrate is stored. Over 3,500 Gaucher Disease patients have been treated since the early 1990s, and the treatment is considered a clinical success (4). The major form of Gaucher Disease does not have CNS involvement. However, the less common neuropathic forms of Gaucher Disease will require a strategy for the enzyme to reach the CNS.

Another apparent success in the treatment of LSD is enzyme therapy for Fabry Disease, another sphingolipid disorder that does not produce lysosomal storage in the CNS (5). This LSD affects primarily vascular endothelial cells and results from a deficiency of α-galactosidase A, which leads to the pathological accumulation of globotriaosylceramide (GL3) and related glycosphingolipids in these cells. Kidney involvement leads to loss of renal function in the third or fourth decade of life. This disease does not affect brain directly, so enzyme access to brain is not required, but it does eventually lead to cerebral vascular insufficiency because of endothelial damage. Two clinical trials of enzyme produced by two different companies were reported recently (6, 7). Although both appear very promising, long-term data are not yet available. Both products have been approved for clinical use in Europe, and approval for both has been sought in the United States.

The MPS storage disorders are also moving up to the plate for enzyme replacement with clinical trials for MPS I (Hurler Disease, α-l-iduronidase deficiency) already reported (8), trials for MPS II (Hunter Disease, α-l-iduronidate sulfate deficiency) under way, and trials for MPS VI (Maroteaux–Lamy Disease, N-acetylgalactosamine-4-sulfatase deficiency) are just beginning. Although MPS VII (Sly Disease, β-glucuronidase deficiency) may be among the last of these disorders to be treated, it played an important role in the evolution of enzyme replacement therapy for the whole group of LSDs (9). Because it appears unlikely that infused lysosomal enzymes will cross the blood–brain barrier, there is not much optimism that i.v. administered enzyme alone will correct the CNS storage present in most of the MPS disorders.

Like the MPS, most of the other LSDs also have CNS involvement and therefore require a strategy for getting enzyme beyond the blood–brain barrier to achieve correction. A number of groups using a variety of viral vectors and enzyme-producing cells have achieved expression of enzyme in brain of animal models (10–17). Impressive degrees of clearing of local, and in some cases distant, storage have been demonstrated. These experiments raised hopes that arresting progression of CNS pathology was possible through brain-directed gene therapy. What few dared to hope was that this approach would not only prevent progression but also erase neurologic deficits. That is just what Davidson's group reports (1).

The authors (1) show that established CNS storage and the related functional deficits in MPS VII mice can be ameliorated by viral-mediated gene therapy. The lentivirus feline immunodeficiency viral vector they used transduced terminally differentiated cells in the brain and mediated β-glucuronidase (GUSB) gene transfer into CNS cells in adult MPS VII mice. This treatment resulted in secretion of GUSB from transduced cells and uptake by nontransduced cells, leading to reduction in preexisting established brain LSD. Correlating with the reduction in storage in the CNS, these adult mice with established behavioral abnormalities related to the lysosomal storage had dramatic improvement in spatial learning and memory when GUSB was expressed. Finally, the correction of the pathology and cognitive improvement were accompanied by changes in expression of genes that have been associated with neuronal plasticity. These observations are particularly important because, as the authors point out, most patients with LSD are not diagnosed until they have established lysosomal storage lesions and functional defects. Recovery of function rather than protection from disease onset is a key goal for any effective therapy for human LSD, because most patients are diagnosed well after onset of CNS disease.

The model the authors (1) chose to study is murine MPS VII, the mouse model for human MPS VII or Sly Disease, which results from GUSB deficiency (18). MPS VII is one of the rarest of the human mucopolysaccharide storage disorders, each of which is produced by deficiency in one of the enzymes involved in the degradation of glycosaminoglycans (GAGs), formerly called mucopolysaccharides. Its importance in the evolution of our understanding of lysosomal enzyme targeting outweighs its clinical significance. When first discovered in the early 1970s, MPS VII had one unique feature among the MPS disorders: the deficient enzyme, GUSB, had been purified and characterized several years before the disease was identified. Addition of GUSB was shown to prevent and correct the accumulation of GAGs in fibroblasts from MPS VII patients (19, 20). Thus, MPS VII immediately attracted attention as a model to study enzyme replacement therapy. Studies of the cultured skin fibroblast model system led to the discovery that uptake of GUSB depends on cell surface receptors that recognize phosphate-containing sugar moieties (Man6-P) on the enzyme (21). The Man6-P residues are added to the GUSB and other acid hydrolases as a means of targeting intracellular enzymes to lysosomes. Another receptor was identified when injected GUSB, from which phosphate had been removed, was found to be rapidly taken up by fixed-tissue macrophage receptors that recognize exposed mannose residues (22). These studies paved the way for Brady and associates to develop “mannose-targeted” cerebrosidase for the treatment of Gaucher Disease (3, 4). These early findings naturally heightened hopes that enzyme replacement in MPS VII patients might lead to correction of lysosomal storage lesions in this disorder. However, MPS VII proved to be too rare (fewer than 100 cases recognized) and too variable to allow controlled experiments to evaluate the response to enzyme therapy.

Nonetheless, hopes for therapy for this and related disorders were greatly advanced by the discovery by Birkenmeier et al. at The Jackson Laboratory that GUSB deficiency in mice produces a disorder resembling Sly Disease in humans (23). (Fig. 1) MPS VII mice have a degenerative disease with progressive disability that reduces life span from an average of 28 to just 5 months. Progressive accumulation of undegraded glycosaminoglycans in lysosomes affects the spleen, liver, kidney, cornea, brain, heart valves, and skeletal system and produces widespread organ dysfunction. Progressive hearing loss leads to early deafness, and defects in learning and memory are evident (24–26).

Figure 1.

An adult MPS VII (Sly Disease) mouse (Left) is much smaller than its phenotypically normal littermate (Right) and has facial dysmorphism with a broad shortened nose and short limbs.

The MPS VII mouse, with a well characterized and uniform genetic constitution and a relatively short life span, proved an attractive model to study experimental therapies for LSD. Mice with MPS VII responded well to bone marrow transplantation, although there was little reduction in brain storage vesicles (27, 28). Enzyme replacement using recombinant GUSB elicited dramatic improvements in visceral pathology but little change in the lyosomal storage lesions in brain unless given to newborns (29–32). Infused GUSB did not cross the blood–brain barrier in mice after 2 wk of age (32). Many promising studies have been reported recently by using this model to study gene therapy, including therapy for CNS storage (10–17). Brain-directed gene therapy, in which viral vectors were introduced directly into the brain, proved one way to bypass the blood–brain barrier, and several studies showed evidence of clearance of CNS storage. However, until now, none of these studies addressed the question raised by Brooks et al. (1): whether therapy that corrected the typical cellular pathology in brain could also erase preexisting neurological deficits. For this reason, the study by Davidson's group represents a major advance in this area.

Given the rapidly expanding number of animal models of LSD with CNS involvement and the generality of the biology of lysosomal enzyme transport, these studies are likely to be replicated in other animal models. If the results in other animal models turn out to be as promising as those presented for murine MPS VII, this study will likely be viewed as a landmark that took us well beyond the blood–brain barrier.