Daily digest — 2026-10-26

Would I take this forward? A conserved channel with atrial-skewed expression is a reasonable start, but expression isn't native current, and ketodoxapram, a doxapram derivative, isn't a clean probe. Before crediting any rhythm effect to TASK1, I'd want its selectivity against other K2P and cardiac channels. With the in vivo readout cut off, the horse-as-AF-model case stays open.

Bile acid shifts can travel with a metabolic benefit without causing it, so I'd look for a knockout comparison before trusting the mechanism. Mouse bile acid pools also differ from ours, which means this model may flatter the human case. The selectivity panel and liver exposure behind the efficacy are the other things to ask for.

Sustained unbound brain exposure after oral dosing is the part I'd trust first, since it lets the cognitive benefit be read against actual target occupancy rather than assumed. Still open: human genetic support for blocking CHI3L1–RAGE, and whether the in vivo benefit tracks engagement of CHI3L1 itself. An aggressive amyloid mouse is a forgiving model, and this is an unreviewed preprint.

The benefit that outruns what contraction predicts is the most interesting result here, and the least secure. In a fast, fragment-transgene model like R6/1 it could just as easily reflect nickase or AAV expression effects, so a catalytically dead control and a full-length knock-in would settle more than another efficacy readout. Long-term nick-related toxicity and spread beyond the striatum also need data.

A glowing tumor is not yet proof of target engagement. Tracking with expression is encouraging, but a blocking-dose or dye-only control would show whether the signal reflects ChoKα binding or a lipophilic dye pooling in tissue. Spontaneous canine tumors are a far better test than mouse models; for people, injection timing and false positives at the margin remain unresolved.

The same depletion that seems to buy less chronic GVHD also strips early pathogen coverage, so a narrower repertoire reads as a trade-off, not a win. This is a human cohort readout: association, not mechanism. For anyone designing a prophylaxis combination, the prize is sparing protective clones while pruning alloreactive ones, and nothing here shows that separation can be achieved.

Lower vesicle CRKII could be a mechanism of apatinib or just a downstream readout of VEGFR2 blockade. Knockdown and rescue make the link convincing in cells, but it has to hold at clinically reachable exposure. And slower growth with fewer vessels in a syngeneic model says little about recurrence yet; the adjuvant idea remains a hypothesis.

Microsomes can tell you which enzyme makes the psychoactive metabolite; they can't tell you how a person's exposure will split. They give no in vivo exposure to the 11-hydroxy species, and the genotype effect is a prediction, not an observation. Clinical PK in CYP2C9 poor metabolizers, plus 2C9 inhibitor interaction data, would have to come before anyone calls this a risk factor.

Does the PAM's neuroprotection actually need NEGR1? That's the experiment to run first, because a scaffold-dependent mechanism means efficacy could vary with NEGR1 levels across neurons, brain regions and people. Knockdown in cultured primary neurons is also a long way from an intact brain; a rescue experiment and in vivo evidence would have to come before I leaned on it.