AAV-PHP.B fails in primates

The research record

This is the project's file for this landmark, as written and checked. Sources are listed at the end; anything not yet confirmed is marked "Not yet verified".

Who

Juliette Hordeaux, James Wilson and colleagues, University of Pennsylvania (Mol Ther 2018 and 2019). Qin Huang, Ken Chan, Benjamin Deverman and colleagues, Broad Institute (PLoS One 2019). Also Liguore, McBride and colleagues (Mol Ther 2019) and Matsuzaki, Hirai and colleagues (Mol Ther 2019). Funders were not checked; Europe PMC lists NIH support for the Hordeaux 2018, Huang and Liguore papers.

What happened

AAV-PHP.B (landmark 111) reached most mouse brain cells after a vein injection. Gene therapy for Parkinson's would need that in people.

• 2018. Hordeaux et al. (a letter titled "The Neurotropic Properties of AAV-PHP.B Are Limited to C57BL/6J Mice") found the brain boost "was not found in nonhuman primates or the other commonly used mouse strain BALB/cJ". It worked in C57BL/6J mice only. They also reported possible serious acute toxicity in monkeys after high-dose systemic AAV.
• 2019, the receptor. Hordeaux et al. mapped the effect in C57BL/6J mice to a specific Ly6a (Sca-1) gene variant and showed PHP.B binds the LY6A protein. Huang et al. compared 13 mouse strains and called LY6A an essential receptor for the PHP.B family; deleting Ly6a or blocking it cut uptake of the PHP.eB variant by mouse brain endothelial cells, and adding it to HEK293T and CHO cells raised uptake 30-fold or more.
• Rhesus macaques. Liguore et al. found a different pattern than in mice: broad cortical and spinal cord transduction after intrathecal dosing, but very low transduction after intravenous dosing. They saw PHP.B as possibly useful for diseases of the cortex or spinal cord.
• Dissent. Matsuzaki et al. (a letter titled "Neurotropic Properties of AAV-PHP.B Are Shared among Diverse Inbred Strains of Mice", Mol Ther 2019) reported that PHP.B works across diverse inbred mouse strains, so the strain-limited claim is disputed; the primate result was not (their paper has no abstract, so the title is all that was read).

Why it moved us toward a cure

It showed that a delivery tool proven in mice may rest on a mouse quirk. It challenges the enthusiasm of landmark 111 and pushed the field to engineer capsids against human receptors, as in landmark 255. It also explains why current gene therapy trials still inject into the brain (landmark 112).

Outcome

Advanced as a negative result. PHP.B remains a mouse research tool. Whether LY6A has any primate equivalent is not established by the abstracts read. No correction or retraction was found for the five papers; the Hordeaux 2018 and 2019 papers each have an accompanying commentary in Mol Ther.

Sources

• Hordeaux et al., Mol Ther 2018;26(3):664–668, DOI 10.1016/j.ymthe.2018.01.018, PMID 29428298
• Hordeaux et al., Mol Ther 2019;27(5):912–921, DOI 10.1016/j.ymthe.2019.02.013, PMID 30819613
• Huang et al., PLoS One 2019;14(11):e0225206, DOI 10.1371/journal.pone.0225206, PMID 31725765
• Liguore et al., Mol Ther 2019, DOI 10.1016/j.ymthe.2019.07.017; Matsuzaki et al., Mol Ther 2019, DOI 10.1016/j.ymthe.2019.02.016 (abstracts via Europe PMC; Liguore, Mol Ther 2019;27(11):2018–2037, PMID 31420242)
• Not yet verified: that LY6A is mouse-specific (the abstracts say only that the effect depends on a mouse variant and is "species- and strain-specific"); primate group sizes; the toxicity details; Matsuzaki 2018 marmoset data (cited in landmark 111).