Letter to the Editor

Main Text

We wish to bring to your attention inaccuracies in the commentary published by Brian W. Bigger in the recent October 2020 issue of Molecular Therapy.¹ The commentary discusses an article published in the same issue by Lahey et al.² The first inaccuracy involves the role played by the human β-hexosaminidase isozymes HexA and HexB in GM2-ganglioside turnover. While it is true the α- and β subunits of human β-hexosaminidase combine to form three different isoforms—HexA (α/β), HexB (β/β), or HexS (α/α)—only HexA is able to hydrolyze GM2 ganglioside in humans. HexB is not involved in GM2 ganglioside turnover in humans since it lacks an α subunit and thus the α-active site needed to hydrolyze the ganglioside.^3,4 Further, elements of both the α- and β subunit (aside from the β subunit active site) are necessary for HexA to correctly interact with the GM2 Activator Protein.^3,5,6 This latter protein correctly presents GM2 ganglioside to the α subunit active site of HexA and is required for ganglioside hydrolysis.

It is a common misconception to assume HexB is needed for GM2 turnover due to the nomenclature that is used to describe the biochemistry of the hexosaminidase system and the associated diseases. Tay-Sachs disease occurs from loss of HexA activity due to mutations in HEXA (which codes for the α subunit), whereas, Sandhoff disease occurs from loss of HexA due to mutations in HEXB (which codes for the β subunit of HexA).³ Thus, mutations in the β subunit cause disease because of their negative effect on HexA formation, not because of a lack of the HexB isozyme.

This misconception led to a second inaccurate comment about our own work on the development of HexM as a therapeutic strategy to treat both Tay-Sachs and Sandhoff disease. Since Dr. Bigger assumed HexB is needed for GM2 turnover in humans, he incorrectly stated that HexM could not be used to treat Sandhoff disease. This is not true. HexM is an engineered hybrid enzyme that combines the critical features of the α- and β subunits of human HexA into a single subunit (μ).⁶ The resulting μ-subunit forms a stable homodimer (HexM) that degrades GM2 in a GM2 Activator Protein-dependent manner both in cellulo and in vivo when delivered to the CNS using an adeno-associated viral vector.6, 7, 8 Indeed, we have shown HexM is able to reduce GM2 levels in both Tay-Sachs and Sandhoff disease mouse models.6, 7, 8 Unfortunately, while Lahey et al.² mention our work on treating Sandhoff mice with HexM, they failed to cite our article.⁸

Additionally, a statement by Lahey et al.² that “the efficiency [of HexM in Sandhoff disease mice] was limited by an immune response” should have been tempered by a comment that, unlike their construct, which was made from mouse α- and β sequences, HexM is constructed using the human sequences,⁶ which would likely cause an increase in antigenicity in treated mice.

Finally, caution is needed when translating results obtained from mouse models of GM2 gangliosidosis to humans. Mice have a metabolic pathway that converts GM2 ganglioside to its asialo derivative, GA2, which can be catabolized by HexB. This “shunt” pathway bypasses HexA and can thus rescue the phenotype of HexA-deficient mice using HexB alone.⁹ Humans lack this shunt pathway. With this in mind, predicting the therapeutic benefit of a gene therapy in humans must discount the benefit observed in mouse models that is associated with HexB activity.

We appreciate the efforts of Lahey et al.² to advance a therapeutic approach to treating these devastating genetic diseases. However, we felt it necessary to correct the published record on the biochemistry of human hexosaminidase isozymes that are directing important efforts now underway to develop gene-based treatments for Tay-Sachs and Sandhoff disease.