Twenty items tonight: seven medchem (covalent pan-FGFR, NLRP3 and p38α–MK2 glue work), four pharmacology, six computational (a trial-outcome model, synthesis-aware generation) and three industry stories, led by pelacarsen's Lp(a) miss and Novo's ziltivekimab failure at AHA.
2026-11-07 · ⚗️p38α–MK2 molecular glue · Journal of Medicinal Chemistry
- Glue-type inhibitors that bridge p38α and MK2 act longer than classic active-site blockers, but the ternary-complex mechanism makes rational design hard.
- The team combined free energy perturbation with machine learning to rank compounds drawn from varied chemical sources in silico.
- This yielded several structurally distinct, potent, selective glues; two leads showed good exposure and cut TNFα in an LPS mouse model after oral dosing.
💡 PointsBased on the abstractA team would next point this FEP-plus-ML triage at its own induced-proximity target, but first ask for the missing control: how did the ranking beat plain docking or a similarity screen on hit rate? For glues I'd also want cooperativity (ternary versus binary affinity) and residence time reported, since those, not potency alone, explain the durable effect.
Diverse Computational Strategies Enable the Discovery of p38α−MK2 Molecular Glues ↗
2026-11-07 · ⚗️FGFR1-4 covalent kinase inhibitor · Journal of Medicinal Chemistry
- Structure-guided design produced a 3-(aminobenzylidene)indolin-2-one series that binds FGFR1-4 irreversibly, aimed at resistance to approved FGFR drugs.
- The lead compound 9v inhibits all four isoforms at single-digit nanomolar levels and suppresses FGFR-dependent cancer cell growth even more strongly.
- It keeps sub-nanomolar activity on gatekeeper and molecular-brake mutants; mass spectrometry confirmed covalent labeling of FGFR3.
- Tumor growth was clearly reduced in a bladder cancer xenograft.
💡 PointsBased on the abstractWhich cysteine is 9v actually labeling, and what else does it label? Retained potency on gatekeeper and brake mutants is encouraging, but a covalent warhead invites off-target cysteine reactivity. I'd want a cysteine-to-serine mutant control, kinome-wide profiling and glutathione reactivity data, plus a head-to-head against futibatinib, before trusting the selectivity story.
Design, Synthesis, and Biological Evaluation of Indolin-2-one Derivatives as Novel Covalent Pan-FGFR Inhibitors for Overcoming Clinical Resistance ↗
2026-11-07 · ⚗️Transthyretin (TTR) tetramer stabilizer · Journal of Medicinal Chemistry
- A kinase-inhibitor scaffold was repurposed to hold the transthyretin tetramer together, using simplification of the core, tweaks to the terminal phenyl and halogen placement.
- Calorimetry showed gains came mainly from enthalpy, with binding free energy shifting only slightly.
- X-ray and neutron structures point to better shape fit and weak contacts, not classic hydrogen bonds; the lead retains strong binding in human plasma and rivals clinical stabilizers in blocking aggregation.
💡 PointsBased on the abstractEnthalpy works as a design handle only if the gains survive enthalpy-entropy compensation, and a nearly flat free energy is exactly where that bites. I'd want matched pairs isolating the halogen effect, plus data on common destabilizing variants and selectivity over other plasma thyroxine binders, before calling this a better chemotype than the existing benzoxazole-type stabilizers.
Enthalpy-Informed Design of Dasatinib-Derived Small-Molecule Inhibitors of Transthyretin Amyloidogenesis ↗
2026-11-07 · ⚗️NLRP3 inflammasome inhibitor (small molecule) · Journal of Medicinal Chemistry
- Tuning the lipophilic and solvent-facing portions of the sulfonylurea MCC950 gave F34, which keeps nanomolar cellular potency and spares the NLRC4 and AIM2 inflammasomes.
- Docking and molecular dynamics place F34 in the MCC950 site; SPR shows direct, micromolar binding to NLRP3.
- In mice F34 has a long half-life and very high oral bioavailability, and outperforms MCC950 in gout and chronic colitis models, with better tolerability and low hERG and CYP liabilities.
💡 PointsBased on the abstractCompared with MCC950 and the newer sulfonylurea-free NLRP3 chemotypes, F34 is a conservative edit: the acidic sulfonylurea core, and its liver-safety question, is still there. Docking plus SPR is thin for a binding-mode claim, so I'd want a cryo-EM or co-crystal structure with NLRP3, and chronic-dosing liver data, before believing the improved tolerability.
Design, Synthesis, and Evaluation of a Potent NLRP3 Inflammasome Inhibitor with Favorable Pharmacokinetic and Safety Profiles for Chronic Inflammatory Diseases ↗
2026-11-07 · ⚗️Caspase-2 / peptidomimetic aldehyde inhibitors · Journal of Medicinal Chemistry
- A 48-member peptidomimetic aldehyde library was built around P2/P3 modifications, and two leads reached low-nanomolar caspase-2 potency with strong selectivity against caspase-3 and other family members.
- The compounds suppressed tau cleavage and formation of the toxic Δtau314 fragment; co-crystal structures rationalized the SAR.
- A reference compound showed good ADME/PK, yet poor permeability, limited brain exposure and an unstable P5 indoline block progression.
💡 PointsBased on the abstractAn aldehyde warhead on a peptidic scaffold with a labile indoline at P5 is a hard sell for CNS work, and the paper more or less concedes it. Aldehydes hydrate, epimerize and clear fast. Next step: replace the P5 block with a more robust cap, test a less promiscuous reversible warhead such as a nitrile, and check unbound brain exposure before trusting any tau readout.
Potent and Selective Caspase-2 Inhibitors: Biological Profiling, Structural Characterization, and Challenges of Molecular Instability ↗
2026-11-06 · ⚗️siRNA · Journal of Medicinal Chemistry
- Four phosphoramidite building blocks lacking a nucleobase, built on pyrrolidine and aminocyclopentane cores, were prepared for placing into siRNA strands.
- Sense-strand placement loosened the duplex depending on position; central or terminal sites boosted silencing, with the (S)-cyclopentane isomer best.
- Putting the (R)-cyclopentane isomer in the seed region curbed miRNA-like off-targeting without losing on-target effect, including in mice.
- The (S) unit at the 3' end also improved nuclease stability and liver exposure, and doubled as a handle for conjugation.
💡 PointsBased on the abstractUnlike GNA or UNA, which destabilize the duplex mainly through added flexibility in the seed, these amino abasic units also carry a conjugation handle, and stereochemistry seems to separate potency from off-target behavior. That (R)/(S) split is intriguing, but I'd want melting data, Ago2 loading, and a head-to-head against GNA at the same seed position before calling it an upgrade.
Synthesis of Amino-Functionalized Abasic Monomers as Versatile Building Blocks for siRNA Modification ↗
2026-11-01 · ⚗️Sigma receptors (S1R/S2R) and proteasome; bifunctional conjugates · European Journal of Medicinal Chemistry
- Four hybrid molecules join a sigma receptor modulator (RC-106) to a proteasome-directed fragment (IP24) through PEG spacers, testing different attachment points and linker lengths.
- All kept sigma receptor binding; where the linker was attached on the RC-106 core steered S1R versus S2R preference.
- Hybrids outperformed their parents in glioblastoma and myeloma lines, with the best one active at low-micromolar level and also in 3D spheroids.
- They did not block purified 20S proteasome but impaired proteasome function in intact cells, suggesting an indirect, context-dependent mechanism.
💡 PointsBased on the abstractWhere is the unconjugated 1:1 mixture of RC-106 and IP24? Without it, I can't tell whether the hybrids' edge over the parents reflects true bifunctionality or just two lipophilic fragments adding up. Since they spare purified 20S yet impair proteasome function in cells, I'd also want a sigma-dead analogue with the same PEG linker to show receptor engagement actually drives the effect.
First Conjugates based on a Sigma Receptor Modulator and a Proteasome-Impairing Agent. Design, Synthesis and Exploratory Biological Evaluation in Cancer Cells. ↗
2026-11-05 · 🧬Tau aggregation / SUMOylation (Jade1S PHD zinc finger protein) · bioRxiv
- Jade1S, a PHD zinc finger protein at a tauopathy risk locus, binds 0N4R tau through several redundant regions and acts as a SUMO E3 ligase on tau in vitro.
- In cells, tau and Jade1S stabilize each other, while Jade1S preferentially speeds clearance of insoluble tau and leaves the soluble monomer pool largely unchanged.
- SUMO2/3-modified tau oligomers are diverted from autophagy-lysosome handling toward proteasomal degradation, and Jade1S also shows chaperone activity against induced tau oligomerization in vitro.
- Unreviewed preprint using overexpression and in vitro systems; disease relevance remains untested.
💡 PointsBased on the abstractSelective clearance of insoluble tau without touching the soluble monomer is the profile you'd want, but I wouldn't move this forward on overexpression data alone. Mutual stabilization of tau and Jade1S could just as easily trap both in aggregates. The next step is a SUMO-site mutant tau and ligase-dead Jade1S in a seeded neuron model, with endogenous-level expression.
Jade1S SUMOylates tau and targets oligomers for degradation ↗
2026-11-04 · 🧬Public TCRs against cancer peptide-HLA targets (HPV, mutant KRAS, CT83); TCR-based therapy · Science Translational Medicine
- Statistical mining of routine clinical tumor RNA-seq from roughly 150,000 patients found recurrent, shared TCR chains linked to known cancer antigens, including HPV, mutant KRAS and CT83 presented on HLA.
- The authors developed ways to pair enriched alpha and beta chains into full receptors, then tested candidates for antigen recognition.
- Dozens of TCRs across nine peptide-HLA targets were validated, some with strong cytotoxic activity, suggesting public anticancer TCRs are common and a new source for TCR-based therapeutics.
💡 PointsBased on the abstractPublic TCRs are attractive because one receptor could serve many patients, but the HLA restriction still caps who qualifies. What I'd want before taking any of these forward is cross-reactivity screening against a normal-peptide panel and primary tissue, since a strong in vitro kill says little about off-target risk. Mutant KRAS and CT83 differ sharply here, and in vitro cytotoxicity won't tell you which is safer.
Cancer-specific T cell receptor discovery using 150,000 clinical genomics samples ↗
2026-11-03 · 🧬Natural flavonoid; mitophagy-ferroptosis axis · Biochemical Pharmacology
- Tilianin, a flavonoid from the Uyghur herb Dracocephalum moldavica, was tested against doxorubicin cardiotoxicity in H9c2 cardiomyocytes and mice.
- Transcriptomics plus mitochondrial modulators and ferrostatin-1 indicated that doxorubicin drives excessive mitophagy, mitochondrial iron release and lipid peroxidation, culminating in ferroptosis.
- Tilianin dampened this mitophagy, partly restored iron balance, and improved cardiac function and reduced myocardial injury, supporting a cardioprotective adjunct for chemotherapy.
💡 PointsBased on the abstractThe mechanism is tidy, but doxorubicin protection is only worth pursuing if the antitumor effect survives. Iron chelation with dexrazoxane is the benchmark, and it has its own debate over this very issue. I'd want tumor-bearing mice with tilianin on board, a head-to-head against dexrazoxane, and plasma exposure data, since flavonoids often have poor oral bioavailability.
Tilianin counteracts doxorubicin-induced myocardial injury by downregulating mitophagy-dependent ferroptosis ↗
2026-11-01 · 🧬Sepiapterin (oral BH4 precursor) for phenylketonuria · Drug Metabolism and Disposition
- Sepiapterin, a prodrug-like precursor of the cofactor BH4 recently approved for phenylketonuria, is orally available and converted to BH4 quickly and extensively in every species tested.
- Radiolabel recovery was high in rats but low in monkeys and humans, plausibly due to volatile gut-derived metabolites and incorporation of the compound into endogenous biomolecules.
- Metabolism was extensive and metabolite profiles were broadly similar across rat, monkey and human, via reduction, oxidation, methylation and side-chain cleavage.
💡 PointsBased on the abstractLow radiolabel recovery in monkey and human is the red flag I'd chase before trusting the rat package. If label ends up in endogenous pools, total radioactivity can't be read as drug-related exposure, so the rat safety margin may not map cleanly. Next step: a parallel dose with a differently positioned label, plus BH4 and metabolite levels in plasma.
Absorption, distribution, metabolism and excretion of sepiapterin, a natural precursor of tetrahydrobiopterin ↗
2026-11-07 · 💻Trial-outcome prediction / mechanism-based ML · Journal of Chemical Information and Modeling
- Many trial-outcome models lean on trial design, development history or compound identity and so memorize compounds; MERIT uses only molecular and disease features, linking intended and off-target effects through drug–protein, protein–metabolite and immune interaction maps to tissue-specific efficacy and safety.
- On a structure-holdout split over hundreds of small-molecule drugs and thousands of trials, it beat two published methods retrained on the same folds.
- It ranked the approved indication above the failed one for most of a preselected set of drugs that failed one indication but are approved for another, and the authors registered locked, outcome-blind predictions for ongoing phase III drug–indication pairs.
💡 PointsBased on the abstractDropping compound identity and trial design is the right fix for leakage, and a structure-holdout split is a harder test than most. But an overall AUROC is not the question a portfolio team asks; the failed-here, approved-there ranking only matters if it holds on non-obvious cases, so I'd read those 60 compounds one by one. The locked phase III predictions are the real test, and I'd wait for those readouts before this goes near a go/no-go discussion.
MERIT: Mechanism-Driven Model Predicts Drug Outcomes and Nominates Indications for Failed Drugs ↗
2026-11-06 · 💻Low-data virtual screening / UMAP similarity enrichment · Journal of Cheminformatics
- For orphan or understudied targets, MolFuSE embeds compounds with UMAP in a similarity space built from ligands known to bind proteins with the same molecular function, needing no target-specific training data.
- Across nine targets, UMAP embeddings of Mordred 2D descriptors gave better early enrichment of high-potency inhibitors than ECFP4 fingerprints, although ECFP4 ranked actives better overall.
- Performance peaked at a reference set of roughly ten thousand compounds and fell with larger sets, as structurally diverse weak binders dilute the descriptor space.
💡 PointsBased on the abstractThe reference-set sweet spot is the honest finding, and it means curation is the method: somebody has to choose that ten thousand, and that is where bias creeps in. Early enrichment versus global ranking is a real split, so on an orphan target I'd run this next to a plain ECFP similarity search and take the union. First check that the 'potency filter' survives when the reference set shares few chemotypes with the true actives.
Molecular function-guided unsupervised similarity-based enrichment for low-data virtual screening ↗
2026-11-06 · 💻Reaction representation / template applicability prefilter · Journal of Cheminformatics
- Reactions are represented as the product-minus-substrate difference of counted ECFPs; for ECFP-compatible templates, applying the reaction shifts the fingerprint by a constant, mapping graph transformations to translations in fingerprint space.
- Reaction-center ECFPs encode the local environment a template needs, giving a fast coordinate-wise prefilter for whether a template can be applied.
- On USPTO-50k and MetaNetX reactions across ECFP radii, both representations carry information for reaction-class and enzyme-class classification, and reaction-center ECFPs add information beyond reaction ECFPs.
💡 PointsBased on the abstractThe practical value is the cheap applicability prefilter: template-based planners and pathway-search engines spend much of their time on templates that cannot match, and this prunes them before subgraph matching. USPTO-50k classification is a saturated benchmark, so I'd read little into it. Drop it into an existing template-based planner and measure the false-reject rate, because a prefilter that silently discards valid templates costs routes.
Representing chemical and enzymatic reactions in fingerprint space for applicability filtering and classification ↗
2026-10-24 · 💻Structure-based hit-to-lead / fragment-contribution analysis (Notum) · Journal of Computer-Aided Molecular Design
- A decision-support algorithm combines structure-based design with reverse fragment-based design, scoring each fragment's contribution to binding energy and its group efficiency to recommend where to grow a hit.
- Applied retrospectively to a Notum carboxylesterase inhibitor series, its prioritized growth vectors correlated well with the optimization strategies actually pursued experimentally.
- The stated aim is to cut the variability of expert judgment and wasted effort at the hit-to-lead stage.
💡 PointsBased on the abstractAgreement with the published route on one series is weak evidence: those routes are the survivors, and Notum is a crystal-friendly, well-behaved target. I'd ask what it recommends for the vectors that were tried and failed, and whether per-fragment energy decomposition, which is noisy, can separate neighbouring vectors. Good as a prompt for the design meeting, not a substitute for it.
A novel approach for recommendation-based prioritization of growth vectors in hit-to-lead optimization ↗
2026-10-19 · 💻Generative design / synthesis-aware molecule generation · Digital Discovery
- A global search over chemical space optimizes molecular properties while sequential forward-reaction prediction builds each proposal, so the output is a molecule plus a synthetic route.
- Routes start from purchasable building blocks, tying generation directly to what a chemist can order and run rather than to a post-hoc synthetic-accessibility score.
💡 PointsBased on the abstractGenerating the route with the molecule beats scoring synthetic accessibility afterwards, but the route is only as trustworthy as the forward-reaction predictor, and those still produce plausible-looking steps that fail at the bench. Before trusting it I'd have a chemist review routes for a handful of top designs, and run it against an SA-filtered generator on identical property objectives.
RxnMol: synthesis-aware molecular generation via reaction-driven global chemical space search ↗
2026-10-15 · 💻Drug–target interaction prediction / cold-start generalization · Journal of Chemical Information and Modeling2026, 66 (20), 13652–13667
- A dual-channel encoder over pretrained drug and protein sequence-model embeddings captures local functional patterns and global context, and a bidirectional cross-modal alignment module links drug functional groups to protein binding regions.
- On four benchmark sets it was competitive in warm-start settings and generalized better in the target cold-start and blind-start scenarios.
- Ablations and structural visualizations are presented as support for the modules and for residue-level interpretability.
💡 PointsBased on the abstractDTI benchmarks are notorious for ligand and protein bias, so the cold-start and blind splits are where to look; I'd want them with similarity-clustered splits and a ligand-only baseline. Attention-style residue maps are suggestive, not evidence of binding mode, so compare them with known co-crystal contacts before using them as pocket hints. For a sparsely studied target family, the practical question is whether the language-model embeddings carry any signal at all.
Cross-Modal Semantic Alignment for Cold-Start Drug–Target Interaction Prediction ↗
2026-11-07 · 🏢trial result · STAT News · Novartis / Ionis Pharmaceuticals
- Novartis presented the full Lp(a)HORIZON results at the AHA meeting, with simultaneous NEJM publication. The antisense drug, licensed from Ionis, sharply lowered Lp(a) but did not reduce cardiovascular events versus placebo over nearly five years.
- The topline miss was disclosed in September; the full data leave the trial leader saying many in the field will be left scratching their heads.
💡 PointsBased on public reportsThis is the first large test of the Lp(a) hypothesis and it did not pay off, which hits the whole class: Amgen's olpasiran and Lilly's lepodisiran are running similar outcomes trials on the same premise. Two things to watch: whether the degree of Lp(a) lowering, not just direction, matters, and whether a subgroup with very high baseline Lp(a) tells a different story.
STAT+: Failure of Novartis Lp(a) drug raises questions about cardiovascular strategy ↗
2026-11-07 · 🏢trial result · STAT News · Novo Nordisk
- Full ZEUS results were discussed at the AHA meeting. The once-monthly IL-6 antibody did not cut major cardiovascular events in patients with atherosclerotic disease, kidney disease and elevated inflammation, even though it lowered IL-6 and hsCRP as intended.
- Serious infections were more frequent on drug, Novo is taking a non-cash impairment, and the heart failure (HERMES) and post-heart-attack (ARTEMIS) trials continue, reading out in the first half of 2027.
💡 PointsBased on public reportsA clean target-engagement miss is the worst case for IL-6: the biomarker moved and outcomes did not, which undercuts hsCRP as a surrogate for patient selection. Colchicine and Novartis's IL-1β precedent keep the inflammation thesis alive, but HERMES and ARTEMIS now carry the whole franchise, and an infection signal is a hard thing to net against a modest benefit.
STAT+: Despite Novo drug's failure, cardiologists say they will still pursue anti-inflammatories ↗
2026-11-06 · 🏢pipeline progress · BioSpace · Moderna / Merck
- Third-quarter revenue rose about 29% to $1.3 billion on global COVID vaccine sales, and full-year growth guidance was lifted to 15–25%. Analysts questioned how much reflects 2027 deliveries shipped early.
- Intismeran, with Merck, showed a 42% lower risk of recurrence or death in melanoma alongside Keytruda; a BLA is planned by year-end with a decision expected in 2027, and a renal cancer Phase 2 is next.
- The norovirus vaccine missed its early Phase 3 bar, so no further cohort will be enrolled this season.
💡 PointsBased on public reportsProfit this quarter is seasonal COVID timing; the real story is that Moderna's identity is shifting to oncology, with intismeran its first potential approval outside respiratory vaccines. Merck shares the economics, and the open question is the manufacturing and reimbursement model for a bespoke product at launch scale.
Moderna enjoys stable vaccine sales thanks to global COVID uptake ↗
