Daily digest — 2026-10-27

Going after a transient folding intermediate instead of the native fold is a clever way around Cyclin D1's lack of a classic pocket, and it yields a degrader-like outcome by a different route. My worry is selectivity: other cyclins may share intermediate features, and a short-lived state is hard to confirm by SAR. Activity in CDK4/6i-resistant models is the interesting bit; I'd want compound structures and potency first.

Triazolopyrimidine looks like a fairly unusual GCS chemotype next to the ceramide-mimetic and amine-heavy designs I'm used to, so scaffold novelty may be real, and a flat aromatic core is easy to decorate. But "highly potent" with no numbers, and nothing on selectivity, means I can't rank it against eliglustat-type or venglustat-type brain-penetrant compounds. These are patent claims, not peer-reviewed data.

Re-engineering mefloquine's amino-alcohol quinoline is a sensible start, but a lipophilic CF3 plus a basic piperazine is exactly the profile that drives membrane-active, cationic-amphiphilic cytotoxicity, so I'd discount any potency gain until the selectivity window is clear. The (R)-4c enantiomer preference hints at a real target interaction, yet additive rather than synergistic combinations are a thin case for adding it to a regimen.

Three unrelated chemotypes converging on one MYST-family acetyltransferase is the interesting part: acylsulfonylhydrazides, thiadiazines and benzisoxazole sulfonamides plausibly all sit in the acetyl-CoA site, so selectivity over the other MYST members is the real SAR battleground, not potency. Prifetrastat reaching Phase III suggests this polar, sulfonamide-heavy space can be made drug-like.

Swapping NHOH for NHNH2 is a neat fix: it keeps the H-bond donor/acceptor pattern (and the extra Gly86 hinge contact) while dodging the labile, toxic hydroxylamine. But hydrazines carry their own liabilities (reactivity, MAO/aldehyde adducts), so cap or cyclize early. No neuroprotection despite on-target engagement means CK1δ potency alone won't carry this.

Classic HTS trap: when two scaffold classes lose activity on resynthesis, I'd bet on a trace reactive or metal-laden impurity, not a real SAR cliff. Entry-inhibitor chemotypes also tend to be lipophilic and aggregation-prone, so I'd treat any new entry hit as suspect until resynthesized material and an orthogonal assay agree.

Four cysteines, four very different risk profiles. Cys481 in BTK and Cys797 in EGFR are well-validated handles, but the C797S escape mutation shows the weakness: lose the nucleophile and the whole warhead strategy collapses, so reversible or allosteric backups matter. JAK3's Cys909 is the interesting one, since it can buy isoform selectivity that ATP-site binders struggle with, though acrylamide reactivity across the kinome remains the tox question.

Another acrylamide-on-switch-II chemotype, and quinazoline cores are well-trodden ground after adagrasib's tetrahydropyridopyrimidine, so novelty rests on the 6-carbonitrile. Fast covalent engagement plus pERK/pAKT knockdown is the easy part; partial H358 tumor suppression despite decent oral exposure suggests in vivo potency or residence trails sotorasib-class leads. Moderate microsomal stability is fixable, but the abstract says nothing on selectivity against other cysteine-bearing proteins.

HDAC11 is the one isoform where selectivity is cheap to claim and hard to earn, since its fatty-acyl deacylase pocket differs from classic zinc-hydrolase HDACs. If PB151 really clears the other isoforms, that beats most hydroxamate-type tool compounds. But SHMT2 rising is the only engagement readout, and the LKB1-AMPK story looks downstream and possibly off-target. I'd want a cellular thermal shift and an inactive analog first; parenteral dosing also says little about oral exposure.