Splice-switching small molecules and cryptic exon manipulation to lower PrP (CP3 risdiplam derivative; PRNP poison exon)

This approach tries to lower the amount of prion protein (PrP) by changing how the PRNP pre-mRNA is spliced, so that more transcripts include a cryptic exon in the single intron and are then destroyed or poorly translated. Two lines of work exist, both preclinical. A 2024 peer-reviewed genetic study from the Vallabh and Minikel group at the Broad Institute and Massachusetts General Hospital showed that strengthening splice sites around the cryptic exon lowers PrP in cultured cells. A June 2026 preprint from the University of Chicago and the University of Colorado Anschutz Medical Campus reports CP3, a chemically modified version of the approved drug risdiplam, which lowers PrP about 70 percent in human cells and, combined with a Luc7L-raising compound, about 40 percent in the brains of young transgenic mice. It is a small-molecule alternative to the antisense route in approach-prp-aso, and no human data exist.

Key facts

  • Human PRNP has one intron that contains a cryptic exon, homologous to exon 2 of many other mammals; including it would create an upstream open reading frame in the 5-prime untranslated region.
  • In HEK293 cells, a PRNP minigene with strengthened splice sites showed up to 78 percent lower PrP, attributed mainly to repression by the upstream open reading frame, with a possible role for nonsense-mediated decay.
  • The authors of the 2024 paper noted limits: cell culture and a transfected minigene only, no in vivo data, and the strongest variant needed six nucleotide changes, so a drug-like modulator would need a new screen.
  • Risdiplam, an oral SMN2 splicing modifier, was approved by the FDA on 7 August 2020 for spinal muscular atrophy in adults and children 2 months and older, which makes small-molecule splicing modifiers a precedent drug class.
  • CP3 was made by altering risdiplam so it favours weak GU|GU 5-prime splice sites; it promotes a poison exon in PRNP intron 1 that carries a premature stop codon and sends the transcript to nonsense-mediated decay.
  • CP3 was reported as over 60 times more potent than risdiplam on a PRNP minigene, and risdiplam's EC50 at GU|GU sites was about 16 times higher than CP3's.
  • In LN229 human glioblastoma cells, CP3 at 3.2 micromolar cut PRNP mRNA and PrP by about 70 percent.
  • CP3 activity depends on the splicing factor Luc7L; the kinetin analogs PTC258 and BPN-15477 raise Luc7L and let a roughly 5-fold lower CP3 dose reach the same effect.
  • In neonatal Tg26372 mice (human PRNP transgene on a Prnp null background), two intracerebral injections of CP3 plus PTC258 gave about 40 percent lower brain PrP; CP3 alone was shown only in cultured neurons.
  • The preprint was not peer reviewed, reports no survival or disease endpoints, and notes that other genes such as LRRK2 also respond to CP3.
  • The preprint lists funding from NIH, the W. M. Keck Foundation, the CJD Foundation and the Brain Research Foundation, and the authors declare no competing interests.

Timeline

  1. 2020FDA approves risdiplam for spinal muscular atrophy.
  2. 2023Gentile et al. post the cryptic splice site manipulation work on bioRxiv.
  3. 2024The same work is published in J Biol Chem 300:107560.
  4. 2026Liu et al. post the CP3 splice-switching preprint on bioRxiv; later indexed in PubMed and PMC.

People

Last checked 1 November 2026

Sources for this page