VCAP-102 capsid crosses primate blood-brain barrier

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

T. C. Moyer, Mathieu Nonnenmacher (corresponding author) and colleagues at Voyager Therapeutics (Lexington, Massachusetts). Molecular Therapy 33:3902-3916, published online 8 May 2025 (received 14 Dec 2024). The study was funded by Voyager, and several authors are named as inventors on patent applications for the work.

What happened

The first brain-crossing capsid, PHP.B, failed in monkeys because it used a mouse-only receptor (landmark 254). Voyager evolved capsids directly in cynomolgus macaques with its TRACER method (RNA-guided selection of neuron-specific expression), picked 1,500 variants from a peptide library, and tested them in macaques and two mouse strains. The best, VCAP-102, differs from AAV9 by a short inserted peptide.

Given intravenously, VCAP-102 transduced neurons and glia in adult cynomolgus macaques, African green monkeys, marmosets and mice, with 20 to 400-fold higher gene transfer than AAV9 across brain regions and less liver uptake. The range is wide because the species differ. In BALB/c mice, which resist PHP.B, it beat AAV9 by about 20 to 40 times at 10^13 vector genomes per kg. In African green monkeys (same dose) it delivered 4 to 24 times more vector genomes and 16 to 186 times more transgene RNA, and favoured astrocytes over neurons, unlike in mice. In marmosets (2×10^12 vg/kg) it delivered up to 280 times more vector genomes. The receptor is ALPL (alkaline phosphatase); binding to human ALPL started transport across a cell-barrier model.

In June 2026 a Japanese group (Matsuzaki, Hirai) compared eleven brain-crossing capsids in marmosets. Most did not clearly beat AAV9; VCAP-102 did, with GFP found mostly in neurons.

Why it moved us toward a cure

Gene therapies that must reach many neurons, such as GBA1 replacement (landmark 303), need a vein-delivered shell that works in primates. VCAP-102 is a candidate, and a second route to a human receptor beside TfR1 (landmark 255). The field's lesson came from landmark 111: test in primates early.

Outcome

Open. It is animal work only, and no human data or Parkinson's programme using it was found. In a 15 May 2025 company release Voyager's head of gene therapy said two programmes were "advancing towards IND filings this year with a partner"; the part of the release I could read does not tie them to VCAP-102 or to Parkinson's. A different company's engineered capsid, planned for CAP-003 (landmark 303), was withdrawn after a death in a related trial; that capsid is not VCAP-102.

Sources

• Moyer et al., Mol Ther 2025, DOI 10.1016/j.ymthe.2025.04.046 (Europe PMC, PMC12461625)
• Matsuzaki, Konno, Sakamoto, Uchida, Terasaki and Hirai, Mol Ther Nucleic Acids 2026, DOI 10.1016/j.omta.2026.201783
• Paper read in full via Europe PMC (PMC12461625); no correction or retraction listed (PMID 40340250). Voyager release, 15 May 2025 (BioSpace)
• Not yet verified: any earlier preprint; whether Voyager has partnered VCAP-102 for a Parkinson's cargo.