Effective gene therapy in an authentic model of Tay-Sachs-related diseases

Abstract

Tay-Sachs disease is a prototypic neurodegenerative disease. Lysosomal storage of GM2 ganglioside in Tay-Sachs and the related disorder, Sandhoff disease, is caused by deficiency of beta-hexosaminidase A, a heterodimeric protein. Tay-Sachs-related diseases (GM2 gangliosidoses) are incurable, but gene therapy has the potential for widespread correction of the underlying lysosomal defect by means of the secretion-recapture cellular pathway for enzymatic complementation. Sandhoff mice, lacking the beta-subunit of hexosaminidase, manifest many signs of classical human Tay-Sachs disease and, with an acute course, die before 20 weeks of age. We treated Sandhoff mice by stereotaxic intracranial inoculation of recombinant adeno-associated viral vectors encoding the complementing human beta-hexosaminidase alpha and beta subunit genes and elements, including an HIV tat sequence, to enhance protein expression and distribution. Animals survived for >1 year with sustained, widespread, and abundant enzyme delivery in the nervous system. Onset of the disease was delayed with preservation of motor function; inflammation and GM2 ganglioside storage in the brain and spinal cord was reduced. Gene delivery of beta-hexosaminidase A by using adeno-associated viral vectors has realistic potential for treating the human Tay-Sachs-related diseases.

Fig. 1.

β-hexosaminidase activity in murine brain and spinal cord after gene therapy. Red staining indicates β-hexosaminidase product in wild-type control (a) and in rAAV2/2β-transduced Sandhoff mice (c–q), absent in untreated Sandhoff mice (b). rAAV2/2β was stereotaxically inoculated at a single site in the right striatum at 4 weeks and killed at the humane end point of 35 weeks of age. Staining extends from the olfactory bulbs (c) to the spinal cord (n–q). Staining is also observed in ependymal cells in lateral ventricles and the central canal of the spinal cord (arrowheads in e and n, respectively) and in white matter, particularly evident in the pyramidal tract (asterisk in n–q). Reconstruction of a single longitudinal section of spinal cord is depicted in o–q. Sections c–q were obtained from the animal shown in Movie 2, which is published as supporting information on the PNAS web site.

Fig. 2.

Quantification and visualization of glycosphingolipids stored in mouse brain. Glycosphinglolipids (GSLs) in untransduced and transduced Sandhoff mouse brain were analyzed by high-performance thin-layer chromatography (A and B) and electron microscopy (C). (A) GSLs were extracted from a wild type aged 21 weeks (lanes 1, 2, and 13), an untransduced Sandhoff mouse aged 16 weeks (lanes 3, 4 and 14), or the brains of Sandhoff mice transduced with rAAVα+β at a single site in the right striatum (lanes 5–12 and 14–18). Vector was injected at 4 weeks of age; the animals were killed at 16 (lanes 5, 6, and15), 20 (lanes 7, 8, and 16), 24 (lanes 9, 10, and 17), and 30 weeks of age (lanes 11, 12, and 18). Right (lanes 1, 3, 5, 7, 9, and 11) and left cerebrum (lanes 2, 4, 6, 8, 10, and 12) and cerebella (lanes 13–18) were dissected and individually analyzed. Pure GM1, GM2, and GA2 gangliosides and the myelin component, galactocerebroside (Galc), were used as standards (STD). (B) GA2 and GM2 content was quantified densitometrically and is represented as the percentage of the content in the untreated Sandhoff mouse, after correcting for loading differences, by using the internal Galc standard. Storage was diminished in all treated Sandhoff brains but increased progressively with age. (C) Neuronal ultrastructure in brain sections from wild-type (d), untransduced (c), and singly rAAVα+β-transduced Sandhoff mice (a and b). A single striatal injection of viral vector was given at 4 weeks, and the tissue harvested at 16 weeks of age. Neurons in the transduced ipsilateral cerebral cortex had no membranous cytoplasmic cell bodies (b), whereas those in the contralateral cortex (a) were distended by the storage vesicles (arrowheads in a) with distortion of the nuclei, as in untreated Sandhoff animals (c). N, nucleus. (Scale bar: 2 μm.)

Fig. 3.

Relationship between β-hexosaminidase activity, glycosphingolipid storage, and inflammatory cells in the cerebral cortex. Coronal sections from wild type aged 16 weeks (a–d), rAAV2/2α+β-transduced aged 29 weeks (humane end point) (e–h), and untransduced Sandhoff mice aged 17 weeks (i–l) were prepared consecutively. Virus was injected at 4 weeks of age. The β-hexosaminidase reaction product stains red (a and e) and is absent in untransduced Sandhoff mice (i).Glycospingolipid storage, detected by neuronal PAS staining, occurs particularly in layers IV and V of the cerebral cortex of untreated Sandhoff mice (arrowheads in l) but was undetectable in cortex from wild-type (d) or transduced Sandhoff mice (h). Activated microglia/macrophages were recognized by immunostaining of the cell-specific marker, CD68 (b, f, and j), and by binding to isolectin B4 (c, g, and k). No cells of microglia/macrophage lineage were detected in wild-type cortex (b and c), and only a few were seen in transduced Sandhoff mice (arrowheads in f and g). Cerebral cortex from untransduced Sandhoff mice contained numerous activated microglia and macrophages (arrowheads in j and k). The number of neurons staining with PAS and the presence of cells recognized by G. simplicifolia isolectin B₄ (GSIB₄) and CD68 antibodies inversely depended on enzymatic activity.

Fig. 4.

Survival, weight, and neurological function after gene therapy in Sandhoff mice. Weight and rescue of neurological function was assessed in wild-type, untransduced, and transduced Sandhoff mice. Transduced animals were injected at 4 weeks of age. (A) Range of body weights in wild-type mice [blue-gray and pink stippled area for males (n = 6) and females (n = 7), respectively], untransduced (cyan triangles; n = 1) and rAAVα-transduced (dark blue squares; n = 1) Sandhoff males, untransduced (pink squares; n = 5) Sandhoff females, and rAAV2/2β or rAAV2/1β-transduced Sandhoff males at four sites (open dark blue triangles; n = 3). After therapy, Sandhoff mice gained and maintained their weight normally. (B) Effect of therapy on hind-limb movements in WT, untransduced or rAAVα-transduced Sandhoff animals (Sandhoff), and Sandhoff mice after transduction with either rAAV2/2β, rAAV2/1β, or rAAV2/2α+β (Treated Sandhoff). Each dot represents a single animal. Over 120 days, movement frequency declined in Sandhoff mice (P = 0.0108) but, in treated Sandhoff animals, remained indistinguishable from WT (P = 0.1107); limb movements improved significantly after gene therapy (P < 0.0001).