The Generative Brain

Serpin (SERPINA3 and SerpinA3n) inhibition as a non-PrP-targeted anti-prion strategy (ARN1468, SISSA and IIT)

Most prion drug work aims at the prion protein itself, either lowering the normal form or blocking its conversion. This approach aims at something else: SERPINA3, a protease inhibitor (also called alpha-1-antichymotrypsin) whose levels rise sharply in prion-diseased human brain, with the mouse counterpart SerpinA3n rising in infected mice. The idea is that blocking the serpin frees proteases to clear prion aggregates, so the drug never touches the prion protein. The SISSA prion biology laboratory in Trieste, working with the Istituto Italiano di Tecnologia in Genoa, published a compound on this basis in 2022: compound 5, ARN1468, which cut prion load in chronically infected cell lines but had negligible brain exposure in mice, so it was never tested in infected animals. A 2025 study with the Carlo Besta institute then found that mice lacking SerpinA3n developed prion disease from three strains much as normal mice do. The approach is therefore an interesting non-PrP target with a tool compound that cannot be dosed and a genetic result that does not support it in its simplest form. Information only, not medical advice.

Key facts

  • The target idea rests on repeated observations that SERPINA3 and its mouse counterpart SerpinA3n go up in prion disease: in human prion disease brain, in prion-infected mice and earlier in BSE-infected macaques.
  • In 128 human post-mortem frontal cortex samples, SERPINA3 messenger RNA was about 40 times control levels in sporadic CJD, about 39 times in familial CJD and in variant CJD, about 53 times in fatal familial insomnia and about 347 times in iatrogenic CJD, with a milder rise (about 7 times) in Alzheimer disease.
  • The same 2017 study found protein levels roughly six times control in CJD brain by Western blot, staining mostly in surviving neurons and reactive astrocytes, but no SERPINA3 messenger RNA in blood and no significant difference in CSF protein between Alzheimer disease and sporadic CJD.
  • A 2022 survey of serpin genes found SERPINA3 the only serpin raised in both human sporadic CJD cortex (about 54 fold) and early Alzheimer cortex (about 33 fold) and in both mouse models, which is why it was picked as the target rather than other serpins.
  • In the same survey, raising SerpinA3n in cells (conditioned medium from overexpressing cells, or 1 micromolar recombinant protein) increased misfolded prion protein by roughly 40 percent, and knockdown reduced prion load, although the reported size of the reduction is inconsistent between the results and discussion sections.
  • ARN1468 came from a structure-based virtual screen of about 15,000 drug-like compounds against a pocket of SERPINA3; 19 were tested in cells, two hits shared a 5-aminopyrazole core, and a similarity search produced the eight analogues from which compound 5 was chosen.
  • In chronically infected mouse cell lines the compound cut misfolded prion protein by about 60 percent (GT1 cells infected with RML), about 35 percent (GT1 with 22L), about 60 percent (N2a with RML) and close to 85 percent (N2a with 22L), with EC50 values of roughly 6 to 19 micromolar against LD50 values of 34 to 55 micromolar in a six-day toxicity test.
  • Repeated treatment across cell passages progressively cleared the infection, and newly infected cells accumulated prions more slowly under treatment, while normal prion protein levels, its location on the cell and cell viability were unchanged, which the authors use to argue the compound is not working on the prion protein itself.
  • The compound did not slow prion protein aggregation in an RT-QuIC assay, it inhibited the SerpinA3n and chymotrypsin complex only at about 1 mM, and isothermal titration calorimetry gave an apparent binding constant of about 26 micromolar at 1 to 1 stoichiometry, so target engagement is weak by drug standards.
  • Mouse pharmacokinetics gave oral bioavailability of about 26 percent, high plasma clearance and negligible brain levels, and the authors state this is why no efficacy study in prion-infected mice was done.
  • In combination with quinacrine in RML-infected N2a cells the authors report synergy (interaction index 0.19), linking this chemistry to the older repurposed-drug line of work.
  • The 2025 genetic test was negative: SerpinA3n-deficient mice inoculated intracerebrally with RML, 139A and ME7 showed no meaningful change in prion propagation on clinical and neuropathological assessment, and the authors suggest other serpins or molecules compensate.
  • The 2022 paper is SISSA-led (Colini Baldeschi, Zattoni, Vanni, Nikolic, Ferracin, Legname) with the molecular modelling and drug discovery, translational pharmacology and analytical chemistry laboratories of the Istituto Italiano di Tecnologia in Genoa, plus the University of Padua, Forschungszentrum Juelich with RWTH Aachen, and the University of Bologna; funding is listed as SISSA and the Helmholtz Association with no grant numbers.
  • The 2025 knockout paper is a Besta and SISSA collaboration with Moda as first author and Legname as corresponding author, also involving the University of Milan, University of Rijeka and a Trieste company (Eurospital).

Timeline

  1. 2014Barbisin and colleagues report SERPINA3 among the transcripts raised in the brains of BSE-infected cynomolgus macaques, an early pointer to the target.
  2. 2017Vanni and colleagues publish the human survey in Scientific Reports showing differential SERPINA3 overexpression across prion disease types.
  3. 2022Zattoni and colleagues publish the serpin survey in Molecular Neurobiology (online 13 April 2022), naming SERPINA3 a potential therapeutic target.
  4. 2022The ARN1468 paper is published online in the Journal of Medicinal Chemistry on 30 June 2022 (print issue 1 July 2022).
  5. 2025Moda and colleagues publish the SerpinA3n knockout prion strain study online in Neurobiology of Disease on 25 May 2025 (print August 2025).
Sources for this page