Archive · 2026 · November 2026

Daily digest — 2026-11-19

Thursday, November 19, 2026

Tonight's 27 items span medchem, pharmacology, computational and industry. Isomorphic Labs' $5B Series C, Pfizer's up-to-$2.6B in vivo CAR-T deal with Starna and GSK's Nelkarmi approval lead the news; in the science, degraders recur and an audit finds a COX-2 ML screen no better than random.

Medicinal Chemistry

6 articles

ACS Chemical Biology

Fluorinated AFMT analogs shield gut L-dopa from bacteria

  • AFMT, a mechanism-based inhibitor of the gut bacterial decarboxylases that destroy L-dopa, can itself be oxidized by human tyrosine hydroxylase, a host metabolic liability.
  • The authors modified AFMT's aryl ring (difluoro analogs) to dodge the host enzyme while keeping activity on the bacterial target.
  • In vitro, the analogs resisted tyrosine hydroxylase turnover yet still blocked bacterial L-dopa metabolism; they are pitched as starting points for animal studies.
PointsBased on the abstract

Fixing the host liability rather than simply dosing more inhibitor is a smart move; carbidopa and benserazide work because they stay out of the gut-bacteria fight. But this is in vitro only: gut exposure, absorption of the fluorinated analogs and microbiome side effects are unknown. Next I'd want SAR on bacterial decarboxylase versus tyrosine hydroxylase, then animal PK.

Journal of Medicinal Chemistry

Autophagy-tethering degrader clears HBV core protein

  • TK-96 joins a capsid assembly modulator (NVR3-778) to an LC3-binding ligand (ispinesib), a first-in-class autophagy-tethering degrader of HBV core protein.
  • It routes both wild-type and drug-resistant core protein to autophagosomes for lysosomal clearance, rather than only blocking capsid assembly.
  • Antiviral activity in HBV-replicating and infected hepatocyte models far exceeded the parent modulator, and oral dosing in carrier mice lowered viral replication, liver core protein and serum antigens.
PointsBased on the abstract

Beating the parent modulator by a wide margin is the payoff you expect from an ATTEC, but I'd want a non-LC3-binding ispinesib analog as control, plus ATG5/ATG7 knockout, to show the effect is truly autophagy-dependent. Ispinesib is also a kinesin Eg5 inhibitor, so off-target mitotic toxicity worries me. Next step: shorten the linker and try a cleaner LC3 binder.

European Journal of Medicinal Chemistry

Tuning 8-hydroxyquinolines to hit GLI1 in medulloblastoma

  • Starting from a known 8-hydroxyquinoline GLI1 inhibitor, analogues were grown into three sub-pockets of the GLI1 zinc-finger site.
  • Lipophilic ligand efficiency guided the work, so potency gains did not come with excess greasiness.
  • The lead racemate was resolved; the (R) enantiomer bound GLI1 more strongly and lowered GLI1 protein in cells, curbed SHH-type medulloblastoma cell growth and migration, and slowed tumors in a xenograft.
PointsBased on the abstract

Growing a known hit into three sub-pockets while tracking LLE is the right way to run this series, and the tighter (R) enantiomer hints at a real binding mode. But 8-hydroxyquinolines are classic metal chelators, so I'd want on-target proof (a GLI1 zinc-finger mutant or knockdown rescue) and selectivity against other zinc proteins. How does it fare against GANT61 or arsenic trioxide?

Journal of Medicinal Chemistry

RL linker designer yields oral covalent FGFR4 degrader

  • PotentialLinker is a reinforcement-learning tool that optimizes PROTAC linkers, scoring candidates by predicted degradation potential without needing precomputed protein-protein complexes.
  • Retrospective STAT3 and RIPK1 case studies recapitulated known linker SAR trends and pointed to suitable E3 ligases.
  • Prospectively it produced FGFR4 degrader D11, which engages Cys552 covalently, degrades selectively, and shows oral exposure plus tumor suppression in vivo with acceptable tolerability.
PointsBased on the abstract

Covalency on Cys552, a residue unique to FGFR4 within the FGFR family, probably drives selectivity more than the linker does. What's missing is a head-to-head with the non-degrading covalent warhead: if that is equally effective in vivo, the degrader premium is unproven. Oral exposure is rare for PROTACs, so I'd want the full PK.

European Journal of Medicinal Chemistry

Glue degrader clears DNMT1 to ease liver fibrosis

  • A dicyanopyridine-based glue degrader, C39, emerged from structure-guided optimization and removes DNMT1 protein in hepatic stellate cells with good potency and a fairly short half-life.
  • In TGF-β1-activated LX-2 cells it dampened stellate cell activation and matrix accumulation.
  • In a CCl4 mouse model it lowered liver injury, collagen deposition and α-SMA, and reduced pro-fibrotic M2-like macrophages and myeloid-derived suppressor cells.
PointsBased on the abstract

A glue aimed at DNMT1 is a welcome alternative to nucleoside traps: decitabine-type drugs are incorporated into DNA, cause damage and are chemically unstable, whereas a glue works catalytically at the protein level. Before calling the stellate-cell phenotype on-target I'd want the E3/DCAF identity, proteome-wide selectivity (DNMT3A/3B, other methyltransferases) and a non-degrading analog control.

Angewandte Chemie International Edition

One-step photo-probes from drug heterocycles

  • Photochemical Minisci chemistry converts medicinally relevant N-heterocycles into photo-crosslinking probes in a single step, with no pre-installed functional groups or redesigned syntheses.
  • Applied late-stage to existing drug-like cores, it installs the crosslinker together with a handle for isolating conjugates.
  • Pioglitazone probe analogs were made and used in photocrosslinking and pull-down experiments to validate its PPARγ target.
PointsBased on the abstract

Handy for any team with a lead in hand: no linker-and-diazirine route to redo, and the probe keeps the drug's own scaffold. The catch is that Minisci C–H functionalization can give regioisomers, and a bulky crosslinker plus handle near the pharmacophore can cost potency or shift binding mode. Confirm each probe still engages PPARγ, and run competition controls against off-target labeling.

Pharmacology

6 articles

ACS Pharmacology & Translational Science

Adenosine Shifts May Blunt Bacterial Defense in Severe COVID-19 and Sepsis

  • Patient plasma profiling by LC-MS/MS showed disturbed ATP, AMP and adenosine in severe COVID-19 and sepsis; non-survivors showed a steeper ATP-to-adenosine shift, and adenosine was higher in septic non-survivors.
  • Healthy-donor immune cells exposed to severe COVID-19 plasma took on an exhausted T-cell phenotype, and cells exposed to septic ICU plasma cleared E. coli less effectively.
  • Adding adenosine weakened bacterial killing, while the A1 antagonist DPCPX alone improved it, pointing to adenosine receptor signaling as a possible therapeutic handle (abstract-level, in vitro).
PointsBased on the abstract

An A1 antagonist improving killing on its own is the odd result here, and I'd want it explained before anyone calls this a target. DPCPX alone helping implies tonic adenosine tone in the assay, yet plasma adenosine is notoriously unstable ex vivo. Next step: A2A versus A1 selectivity, CD39/CD73 inhibition as a comparator, and a bacterial-burden readout in a septic animal model.

bioRxiv

Colon organoids show how microvilli shape drug uptake into cells

  • Preprint: human colon-derived 3D organoids were built as an in vitro model for colonic drug handling, a gap since most organoid ADME work uses small intestine.
  • Proteomics, imaging, lipidomics and functional assays showed isolated primary epithelial cells are a better native benchmark than whole tissue; the organoids resembled it closely.
  • Intracellular drug availability depended on lipophilicity and microvilli, and colonoids differed from HEK293 cells in membrane lipid content and drug binding.
PointsBased on the abstract

Colon is where modified-release and poorly absorbed drugs actually act, so a colonoid that tracks isolated native epithelium is more useful than yet another ileum-flavored model. But the benchmark is the weak link: no one has shown it predicts human colonic exposure. Test it with a few compounds whose luminal or biopsy concentrations are known, and add efflux-inhibitor controls before trusting the P-gp/BCRP readouts.

Chemical Research in Toxicology

Leaky mitochondria push Alzheimer's astrocytes toward ferroptosis via cGAS-STING

  • Amyloid-beta exposure makes astrocytes release mitochondrial DNA, which switches on cGAS-STING signalling, drives autophagy and ultimately ferroptosis.
  • Blocking the mitochondrial permeability transition pore with cyclosporin A curbed this chain in astrocyte cultures; re-activating autophagy or inducing ferroptosis partly reversed the benefit.
  • In APP/PS1 mice, hippocampal cyclosporin A improved cognition and tissue damage, pointing to mtDNA leakage as a possible therapeutic target (abstract-only source).
PointsBased on the abstract

Cyclosporin A is a poor basis for calling mtDNA leakage a target: it hits cyclophilin D, calcineurin and more, and barely crosses the blood-brain barrier, hence the intrahippocampal dosing. Without genetic mPTP/CypD knockdown or STING knockout in astrocytes, the cGAS-STING link stays correlative. Next step: cell-type-specific STING deletion in APP/PS1, plus human astrocyte or iPSC data.

Biochemical Pharmacology

Poria-derived poricoic acid A restores paclitaxel response in stressed breast cancer mice

  • In mouse breast cancer models, emotional distress sped up tumor growth and blunted the effect of paclitaxel.
  • Poricoic acid A, a triterpenoid from Poria cocos, eased depressive-like behavior and brought paclitaxel sensitivity back; at the high dose, the combination gave the strongest tumor suppression, with no obvious systemic toxicity.
  • The compound appears to act on neuropeptide Y receptor 1, blocking NPY-driven ERK activation and boosting paclitaxel-induced apoptosis, which links distress to chemoresistance through a neuro-tumor axis.
PointsBased on the abstract

Before anyone chases NPY1R here, I'd want proof the compound actually binds it. A crude-ish triterpenoid that also lifts mood-like behavior could be hitting many targets, and a receptor antagonist or knockdown rescue is the missing control. Also ask what plasma and brain exposure the high dose gives, and whether paclitaxel alone in unstressed mice is the right benchmark. Mouse-only, so human NPY1R evidence is the real gate.

Biochemical Pharmacology

Oroxylin A calms pneumonia by curbing macrophage ferroptosis via HSP90α

  • In a mouse pneumonia model, the natural compound Oroxylin A eased weight loss, lung tissue damage and inflammatory signaling; single-cell RNA sequencing pointed to macrophages as the main responding cells.
  • In LPS-stimulated macrophages it dampened cytokine output and raised the ferroptosis-protective factors GPX4 and SLC7A11, an effect blunted by the ferroptosis inducer erastin.
  • Target-profiling, molecular dynamics and thermal-shift assays nominated HSP90α as a likely binding partner, linking it to the ferroptosis-driven inflammation (the abstract is truncated beyond this point).
PointsBased on the abstract

A flavonoid with a plausible HSP90α hit is exactly where I'd demand genetic proof before believing the mechanism. HSP90 is a notoriously promiscuous chaperone, so CETSA shift plus docking doesn't show the phenotype runs through it. I'd want HSP90α knockdown abolishing the benefit, plus a ferroptosis inhibitor rescue in vivo, and lung exposure at the dose used.

Biochemical Pharmacology

Znf667 shields the heart by switching off Pcsk6-driven autophagy

  • In heart-attack patients treated with primary PCI, circulating PCSK6 was higher and tracked with injury markers; it also rose in reperfused mouse myocardium and hypoxia-reperfused cardiomyocytes.
  • Silencing Pcsk6 dampened autophagy and shrank infarct size in mice, while overexpressing it boosted autophagy and lowered cell viability.
  • The transcription factor Znf667, reduced after hypoxia-reperfusion, represses the Pcsk6 promoter; restoring it protected the heart, and this was undone by Pcsk6, with NF-kB signalling implicated.
PointsBased on the abstract

Elevated serum PCSK6 after PCI is a correlation, not proof it drives injury, and autophagy in reperfusion is notoriously double-edged: inhibit it at the wrong time and you lose protective flux. Before chasing Znf667, I'd want autophagic flux assays (with bafilomycin) and timing of Pcsk6 knockdown relative to reperfusion. A transcription factor is also a hard drug target; a Pcsk6 inhibitor is the more tractable test.

Computational & AI

7 articles

chemRxiv

Standard ML virtual screen of COX-2 yields candidates that dock no better than random compounds

  • Random forest potency models for COX-2 and COX-1 were used to screen a large ZINC subset, and filtering left a small candidate list that was docked with Vina alongside a random control set from the same library.
  • Docking was checked by redocking a reference inhibitor and by showing the protocol separates known actives from random compounds.
  • The ML candidates docked no better than random, model predictions did not track docking ranks, and most candidates fell outside the applicability domain, despite good retrospective benchmark scores.
  • The authors argue that applicability-domain coverage and same-library random baselines should be required reporting.
PointsBased on the abstract

A 'same-library random baseline' is the right control, but Vina is a weak judge: its scores barely correlate with measured COX-2 affinity, so 'no better than random' may partly reflect docking noise. The sharper lesson is the domain gap: training sets of known NSAID-like chemotypes versus ZINC. Next step: confirm a handful of top picks in a biochemical assay.

chemRxiv

Pairwise learning on a matched-pair graph predicts which modification boosts cyclic peptide permeability

  • CycPeptMMP predicts the permeability change between two peptides linked in a matched-pair graph, rather than one peptide's absolute value, sidestepping differences between publications.
  • The graph was built for cyclic peptides using four common modification types, covering several thousand peptides from seven literature sources; five GNN backbones and repeated random splits compared pair and single modes.
  • Evaluation was purely retrospective on curated data: pair mode helps most for smaller sources and for single-modification analogues, while single mode wins on larger sources.
  • A censored loss handles clamped measurements, and pair mode ranks such pairs more accurately.
PointsBased on the abstract

Predicting the delta between two close analogues is how med chemists actually reason, and it cancels much of the assay and lab offset that wrecks absolute permeability models across papers. The catch: random splits keep near-identical analogues on both sides, so gains may be optimistic. Worth testing with scaffold or source-held-out splits, then a prospective round.

Journal of Chemical Information and Modeling (ACS)

Curated, bias-aware benchmark shows protein–ligand scoring performance drops versus CASF-2016

  • A new benchmark built from PDB and BindingDB entries uses tiered filtering, manual electron-density checks, affinity verification and property-matched decoys to improve structural reliability and reduce bias.
  • Nine scoring functions were evaluated retrospectively on scoring, ranking, docking and a screening-like decoy task.
  • Most tasks showed weaker results than on CASF-2016, while the decoy task shifted model rankings rather than lowering all scores uniformly.
  • Overlap with training data varied widely between methods, so performance depends on curation, protocol and data exposure.
PointsBased on the abstract

A decoy task that reshuffles rankings rather than just lowering scores is the useful signal: it suggests some functions were leaning on property or pose artifacts that CASF-2016 never penalized. Before trusting any leaderboard, check training overlap per method. Then rerun your shortlisted scorers on our own series with matched decoys and known inactives, and see if the order holds.

chemRxiv

Lipid composition and permeant properties jointly shape passive membrane permeation free energies in MD

  • Adaptive-weight-histogram all-atom MD (MemCross) computed permeation free energy profiles for a small, structurally varied set of drug-like molecules.
  • Bilayers were POPC mixed in turn with cholesterol, POPE, a polyunsaturated PC, or sphingomyelin, mimicking plasma membrane diversity.
  • The work is purely computational; permeation depended on both lipid makeup and permeant properties, and mechanistic descriptors were distilled.
  • Multivariate models built on these descriptors point to quantities obtainable from cheaper unbiased simulations.
PointsBased on the abstract

Six compounds can't support multivariate models, so treat the descriptors as hypotheses, not a predictor. Without a PAMPA or Caco-2 anchor, we also can't tell whether lipid-specific free-energy shifts show up in assays, and transporters and efflux are ignored. Still useful: if unbiased-MD descriptors reproduce these shifts, they could add physics-based features to ML permeability models.

chemRxiv

AI-designed ligands bind NCS-1 and enhance its CB1R interaction, with crystal structures

  • REINVENT4 generation was steered by predicted binding to the NCS-1 hydrophobic crevice, synthetic accessibility and CNS-oriented ADMET, including blood-brain barrier permeability; the outputs guided picking purchasable analogues from public libraries.
  • A small set of selected compounds was tested experimentally: nearly all bound NCS-1, and several showed micromolar affinity.
  • Cell-based in situ PPI assays flagged compounds that enhance the NCS-1/CB1R interaction, described as the first such small molecules; high-resolution crystal structures show how they occupy the crevice.
PointsBased on the abstract

The missing control is the story: nobody shows that REINVENT4 beat a plain similarity search or docking of the same purchasable library, and buying analogues blurs what generation contributed. Micromolar binders are fine as tools, but with selectivity against other NCS family members unknown, the CB1R enhancement can't yet be pinned to NCS-1 engagement. Crystal structures give a solid SAR starting point.

arXiv

RL-generated neuraminidase inhibitors show activity against influenza A and B in cells and mice

  • A reinforcement-learning molecule design platform generated a large, diverse library aimed at influenza A and B neuraminidase, and most designs had better physicochemical profiles than oseltamivir.
  • After computational filtering, a small set of non-sialic-acid-like compounds was tested experimentally; two inhibited both virus types, including an oseltamivir-resistant mutant, and simulations suggested oseltamivir-like residue contacts.
  • Both hits were active in cell assays and gave protection in infected mice, supporting generative RL as a route to antivirals validated in vitro and in vivo.
PointsBased on the abstract

Beating oseltamivir on physchem is a low bar, since it's a polar carboxylate prodrug with modest properties. The real test is whether the hits are truly novel versus known NA chemotypes (zanamivir, peramivir, other transition-state mimics), and whether RL beat a simple enumeration or docking baseline. Next step: enzyme IC50 shifts across H275Y, plus PK and dose-response in vivo.

Journal of Chemical Theory and Computation (ACS)

Water-transfer alchemical method estimates RBFE when a bridging water is displaced

  • Extends the Alchemical Transfer Method so a ligand and its bound water are moved together between the binding site and bulk, handling congeneric pairs that differ in substituent and in the number of bridging waters.
  • This sidesteps the sampling problem of kinetically trapped waters that standard relative binding free energy protocols struggle with.
  • Evaluated retrospectively on a protein-ligand set previously used to test water-displacement and water-sampling methods.
  • Predictions were reported as more accurate, reliable and reproducible than standard approaches.
PointsBased on the abstract

Retrospective results on a set already used to benchmark water methods tell me little about my own series. The real question is how the water to move is chosen: if it needs prior crystallographic or GCMC insight, the method is only as automatic as that call. Versus coupling/decoupling or GCMC, I'd want cost per edge and agreement when the water count changes.

Industry News

8 articles

PfizerBioSpace

Bourla: Chinese biotechs will be Pfizer's main rivals by 2030, and AI is the answer

  • At the Jefferies healthcare conference in London, Pfizer CEO Albert Bourla said that by 2030 his main competitors will be Chinese companies, working at 3-4x the speed and half the cost.
  • He called AI the biggest lever for catching up, across the whole business including sales and marketing.
  • China's five-year plan targets discovering a third of global blockbuster drugs by 2030.
PointsBased on public reports

Bourla's own deals undercut the rhetoric: Pfizer's recent licensing with Innovent, Sciwind and Starna means it is buying the Chinese speed it says it fears. The AI-savings claim is the unproven part, since drug R&D cost is dominated by clinical trials that AI doesn't yet shorten. Expect more big-pharma in-licensing from China, and a push to show internal cost cuts, as the realistic response.

GSK / NuvalentBioSpace

GSK's Nelkarmi wins accelerated approval, second drug from the $10.6B Nuvalent buy

  • FDA granted accelerated approval to Nelkarmi (neladalkib) for previously treated ALK-positive NSCLC, 10 days ahead of its November 27 action date; a confirmatory study is required.
  • Approval rests on the Phase 1/2 ALKOVE-1 study: response rates were higher in patients whose only prior ALK inhibitor was Alecensa, with notable intracranial activity. Liver toxicity was the most common reason for discontinuation.
  • It follows July's Jideytro (zidesamtinib, ROS1) from the same acquisition.
PointsBased on public reports

Two approvals within five months of closing makes the $10.6B Nuvalent price look better than it did in June. But the second-line launch faces Pfizer's Lorbrena, and the real prize is first-line, where the head-to-head against Alecensa is still running; the accelerated label rides on single-arm response data until then. Watch hepatotoxicity in that comparison: it is the first place a clean profile could separate from entrenched rivals.

MerckEndpoints News

Merck stops nemtabrutinib, the BTK inhibitor behind its $2.7B ArQule deal

  • Merck has ended development of nemtabrutinib (MK-1026), the main reason it bought ArQule for $2.7 billion in 2019.
  • The drug was in a slew of trials in blood cancers; Fierce Biotech's headline ties the decision to Phase 3 data.
  • Merck had earlier booked a write-down on the asset.
PointsBased on public reports

This closes the book on one of Merck's weaker acquisitions: a non-covalent BTK inhibitor was meant to answer resistance to covalent drugs, but BTK degraders and strong incumbents narrowed the opening. For Merck, the loss is mostly sunk cost; the more telling point is how little the ArQule price bought beyond this molecule. Competitors in the non-covalent class, like Lilly's pirtobrutinib, face one fewer rival.

Pfizer / Starna TherapeuticsBioSpace

Pfizer buys into in vivo CAR T: Starna licensing deal worth up to $2.6B

  • Pfizer takes exclusive rights outside Greater China to three preclinical in vivo CAR T candidates for autoimmune disease from Starna, a China-based RNA biotech.
  • Terms: $330 million upfront, with up to $2.3 billion more in milestones and royalties; Pfizer also gets the platform for immunology and oncology, plus a research collaboration.
  • Starna delivers RNA through a targeted lipid nanoparticle; Pfizer says it is the company's first in vivo CAR T work.
PointsBased on public reports

The $330M upfront is a real commitment for three preclinical assets, so Pfizer is paying for the delivery platform, not the programs. In vivo CAR T skips manufacturing and lymphodepletion, but it also has to prove that a targeted LNP reaches T cells cleanly in humans, and that is untested. With peers pausing ex vivo autoimmune CAR T trials after safety events, Pfizer gets a late entry with a cheaper option. It is also another Chinese in-license in a year full of them.

White HouseBioSpace

Overton FDA nomination pulled, leaving the agency under an interim chief

  • The White House withdrew Heidi Overton's nomination as FDA commissioner on Tuesday evening, hours before a scheduled Senate HELP Committee vote.
  • Republican senators, Democrats and some MAHA-aligned voices objected, citing her backing of an MMR-splitting executive order and reluctance to rule out advisory committee shutdowns.
  • Kyle Diamantas has been interim commissioner since May, when Marty Makary left.
PointsBased on public reports

RBC calls it bad optics but not an operational threat, and that fits: CDER and CBER now have permanent heads, which matters more for approval timing than the commissioner's chair. The real change is political. A nominee after January 3 faces a Democratic Senate, so a long interim stretch is likely, and sponsors should price in continued policy volatility on vaccines and advisory committees rather than on review speed.

AstraZeneca / AlexionBioPharma Dive

AstraZeneca's Klygefa approved as a self-administered C5 option in myasthenia gravis

  • FDA approved gefurulimab, to be sold as Klygefa, for generalized myasthenia gravis.
  • It is a C5 inhibitor that AstraZeneca gained through its Alexion acquisition.
  • Positioned as self-administered, in contrast to Alexion's infused Soliris and Ultomiris.
PointsBased on public reports

The strategic point is cannibalization management: AstraZeneca is launching a self-injected C5 inhibitor against its own infused franchise, which only makes sense as a defense against FcRn and generic eculizumab pressure. Convenience is the pitch, so pricing relative to Ultomiris and uptake among patients already on infusions are the numbers to watch; the roundup gives no efficacy or price detail, so neither can be judged yet.

Isomorphic LabsBioSpace

Isomorphic's $5B Series C is the largest private raise in biotech history

  • Google-backed Isomorphic Labs announced a $5 billion Series C, described as the largest round in the industry's history.
  • It comes six months after a $2.1 billion Series B in May; the company has raised at least $7.7 billion in total.
  • Isomorphic has not yet entered the clinic or disclosed a pipeline.
PointsBased on public reports

This is a bet on a model, not a portfolio: $7.7B raised with no clinical asset means investors are pricing the platform's ability to out-design the field. It also resets what an AI-bio round can look like and starves smaller AI-discovery startups of attention. The proof point to watch is the first declared program and its partner economics, since big pharma deals so far have paid in biobucks rather than cash.

Biogen / TRIMTECH TherapeuticsFierce Biotech

Biogen backs TRIM21 degraders for neurodegeneration in $411M TrimTech deal

  • Biogen pays TRIMTECH about $32 million upfront for small-molecule degraders aimed at neurodegenerative disease, with up to $411 million in total payments plus royalties.
  • It is a research, option and license structure: Biogen can take a global exclusive license to resulting degraders and then funds development and sales.
  • TRIMTECH uses the E3 ligase TRIM21 to clear toxic protein aggregates, as an alternative to conventional PROTACs; the company was founded in 2023.
PointsBased on public reports

The cash that is actually committed is small, so this is an option on a mechanism, not a pipeline buy. The appeal is TRIM21's ability to act on aggregates that CRBN- or VHL-based PROTACs handle poorly. The open question is brain exposure: bivalent degraders are large, and CNS-penetrant chemistry decides whether Biogen ever exercises. Following its glue pact with Neomorph, Biogen is building a degrader portfolio on cheap options.