Neonatal tolerance induction enables accurate evaluation of gene therapy for MPS I in a canine model

Highlights

• MPS I dogs treated with gene therapy develop immune response against human IDUA.

• MPS I dogs exposed to human IDUA as newborns develop tolerance to the protein.

• Tolerance induction allows for accurate evaluation of human vector in canine model.

• Important approach for preclinical evaluation of gene or protein therapeutics

• Potential to induce tolerance to therapeutic proteins for recessive diseases

Abstract

High fidelity animal models of human disease are essential for preclinical evaluation of novel gene and protein therapeutics. However, these studies can be complicated by exaggerated immune responses against the human transgene. Here we demonstrate that dogs with a genetic deficiency of the enzyme α-l-iduronidase (IDUA), a model of the lysosomal storage disease mucopolysaccharidosis type I (MPS I), can be rendered immunologically tolerant to human IDUA through neonatal exposure to the enzyme. Using MPS I dogs tolerized to human IDUA as neonates, we evaluated intrathecal delivery of an adeno-associated virus serotype 9 vector expressing human IDUA as a therapy for the central nervous system manifestations of MPS I. These studies established the efficacy of the human vector in the canine model, and allowed for estimation of the minimum effective dose, providing key information for the design of first-in-human trials. This approach can facilitate evaluation of human therapeutics in relevant animal models, and may also have clinical applications for the prevention of immune responses to gene and protein replacement therapies.

Introduction

Mucopolysaccharidosis type I (MPS I) is a rare genetic disease caused by mutations in the gene encoding α-l-iduronidase (IDUA), an enzyme required for the catabolism of ubiquitous glycosaminoglycans (GAGs) in the lysosome. IDUA deficiency results in lysosomal accumulation of GAGs in many tissues, leading to a variety of clinical manifestations including bone and joint deformities, corneal clouding, and cardiac valve insufficiency. MPS I patients also frequently experience neurological complications such as communicating hydrocephalus and spinal cord compression [1], [2], [3]. The impact of the disease on cognitive function varies; in the attenuated form of MPS I (Scheie syndrome – MIM #607016 or Hurler-Scheie syndrome – MIM #607015) in which there is residual IDUA activity, cognition is affected in only about one third of patients. In the more common severe form of MPS I (Hurler syndrome – MIM #607014), patients universally exhibit rapid cognitive decline in early childhood [4]. MPS I is currently treated with intravenous (IV) infusion of the recombinant enzyme, which can be internalized by cells from the circulation via mannose 6-phosphate receptor binding [5], [6]. Enzyme replacement improves many disease symptoms, but does not reach the central nervous system (CNS), and therefore has no impact on cognitive function [4]. MPS I can also be treated with hematopoietic stem cell transplantation (HSCT), which provides a constant source of circulating IDUA through enzyme secretion by engrafted donor cells. Unlike enzyme replacement, HSCT can improve cognitive outcomes, apparently due to migration of donor-derived cells across the blood-brain barrier, where they serve as a source of secreted enzyme within the CNS. However, HSCT suffers from numerous complications including graft failure, infection, graft versus host disease, and transplant-associated mortality as high as 20% [7], [8], [9], [10], [11], [12], [13]. After transplant, many patients also exhibit residual cognitive deficits, which may be a consequence of disease progression during the slow engraftment of donor cells in the CNS [14]. These shortcomings leave a significant unmet need for a safe and effective therapy for the CNS manifestations of MPS I.

Gene therapy is a promising alternative to HSCT for the treatment of cognitive decline in MPS I patients. Gene transfer has the potential to induce rapid reconstitution of IDUA in the CNS without the adverse effects of HSCT. Successful targeting of even a small number of cells in the CNS could provide a depot of secreted IDUA in the brain, leading to widespread improvement of storage pathology and potentially improving cognitive function. We previously demonstrated that injection of an adeno-associated virus serotype 9 vector (AAV9) into the cerebrospinal fluid (CSF) can efficiently deliver the IDUA gene to cells throughout the CNS, and that the enzyme secreted by transduced cells mediates global resolution of brain storage lesions [15], [16]. These proof-of concept studies for intrathecal (IT) AAV9 gene therapy relied on two naturally occurring large animal disease models, the MPS I dog and MPS I cat. The use of these models was critical for evaluating the efficacy of the approach, not only because they accurately reproduce the CNS pathology of MPS I, but also because these large animals better reflect the human CNS anatomy and CSF circulation compared with rodent models, allowing for realistic representation of the clinical route of administration and the resulting vector distribution. While these initial experiments in the MPS I dog and cat were carried out using vectors expressing species-specific transgenes, advancing this approach toward human trials necessitated the evaluation of a clinical candidate vector expressing the human IDUA transgene. However, studies of the clinical candidate vector in the MPS I dog model were complicated by an exaggerated immune response to the human transgene product. Building on our previous finding that neonatal gene transfer could induce persistent tolerance to the transgene product, we applied this approach to induce tolerance to human IDUA in MPS I dogs, which subsequently allowed for accurate evaluation of the efficacy of the human vector in this high fidelity model [15].