The Generative Brain

Dominant-negative protein-based gene therapy (Castilla)

Instead of lowering the brain's own prion protein, this approach adds an engineered version of the protein that does the normal job but cannot misfold and gets in the way of the harmful misfolded form, delivered to neurons by AAV. In a grantee report from December 2025 treated mice lived 35 to over 100 percent longer than untreated mice in two prion models, but the data are not peer reviewed, the candidate sequences are not disclosed and tests in human-prion mouse models were still running. No primate or human data. Information only.

Key facts

  • Idea in the lead's words: removing PrP removes what prions need, but PrP has unknown functions, so the group proposes a dominant-negative protein that performs the normal functions, cannot misfold and interferes with misfolded protein.
  • Funding: a 2024 CJD Foundation research grant, in vivo proof-of-concept of dominant negative protein-based gene therapy for different prion diseases, at CIC bioGUNE in Derio, Spain.
  • Candidate selection: more than 900 species screened and 424 recombinant PrP variants tested for resistance to spontaneous misfolding, then for interference with prion conversion.
  • Delivery: AAV with a human synapsin promoter, MVM enhancer and WPRE gave neuron-specific expression at 0.5 to 2.0 times endogenous PrP after intravenous dosing, stable for at least 9 months; direct CNS routes did not give brain-wide spread.
  • Survival, RML prions in mice: 161 days untreated versus 217 days with candidate DN1, a 35 percent extension.
  • Survival, bank-vole-adapted CWD prions: 73 days untreated versus 105 days with DN1 (44 percent) and 160 days with DN2 (over 100 percent, still ongoing when reported).
  • Human-relevant models: CJD MM1 prions in human-PrP mice were under test (200 to 240 days expected); GSS A117V and FFI studies were planned.
  • Structure: NMR shows one lead candidate keeps the native PrP fold, which the report uses to argue against non-specific toxicity.
  • Safety signal stated by the grantee: unexpectedly high expression caused toxicity at certain doses, handled by dose studies and promoter changes.
  • Status of the evidence: both reports are grantee-written, funder-hosted and not peer reviewed; a manuscript on the AAV-PrP model system was under review in December 2025, and a Brain Pathology paper with a related title appears on the lab profile dated January 2026.
  • Related evidence: a February 2026 preprint with Castilla as co-author shows in cell culture that G127V PrP (the human variant that protects against kuru) blocks prion infection across strains and keeps working after its expression stops; it does not say it is the same molecule as DN1 or DN2.
  • Lab context: the Prion Research Lab at CIC bioGUNE, led by Castilla since 2009, lists dominant-negative gene therapy in its treatment research line.

Timeline

  1. 2024Grant page published; 2024 CJD Foundation grant awarded.
  2. 2025First-year report posted, noting toxicity at high expression.
  3. 2025Final grant report: mouse survival extensions of 35 to over 100 percent reported.
  4. 2026Brain Pathology paper on AAV-delivered PrP variants in knockout mice listed on the lab profile (not opened).
  5. 2026Preprint on G127V dominant-negative effect in cell culture co-authored by Castilla.
Sources for this page