NEWS.

In the first head-to-head trial, a self-assembling peptide sprayed onto a fresh ooze after a bile-duct cut sealed it on the first try in 92 percent of patients, against 77 percent for the standard balloon, and needed a backup device a third as often.

MOTS-c, a peptide cells make from their own mitochondrial DNA, preserved leg muscle in mice with cancer cachexia by quieting the genes that take muscle apart. It did not stop the weight or fat loss, and the work is still in mice.

B7-H3 is on many solid tumors and the subject of a wave of antibody-drug conjugates and CAR-T trials, but those only help patients whose tumors carry it. A Fudan team built a gallium-68 bicyclic peptide PET tracer whose signal tracked B7-H3 levels in mice, a step toward imaging who to treat. Preclinical.

This week's headline fix for muscle loss on obesity drugs was an antibody, apitegromab, in Nature Medicine. A day later a University of Alberta mouse study showed an oral ketone ester preserved muscle on semaglutide without blunting fat loss. The cheapest answer yet, and still only preclinical.

Novo Nordisk shut down EVOKE and EVOKE+ when semaglutide missed its week-104 endpoint. A single-author paper re-ran the unanalyzed week-130 and 156 data and found a late signal. The caveats, post-hoc tests on summary statistics by a conflicted author, are heavy.

A de novo macrocyclic peptide blocked the Wnt5a-to-ROR1 handshake and halted DIPG cells, while equally tight-binding relatives had no effect. Cell-line only, but it gives an incurable childhood brainstem tumor its first validated receptor target.

A target trial emulation of 3,572 pregnancies found that continuing a GLP-1 drug into the first trimester did not clearly raise the risk of pregnancy loss, growth problems, or birth defects, but the confidence intervals were wide enough that a real increase could not be ruled out.

A meta-analysis of four trials and 655 patients found CGRP monoclonal antibodies cut cluster headache attacks by 0.81 per week, an effect too small and too uncertain to call, with a placebo response between 27 and 53 percent doing much of the work.

A target trial emulation of 161,798 obese, nondiabetic adults found GLP-1 users had a hazard ratio of 0.59 for obesity-associated cancers, about 41 percent fewer diagnoses. The protection held in every subgroup but one.

Innovent's mazdutide beat semaglutide in a head-to-head diabetes trial, with more than twice as many patients hitting both their blood-sugar and weight goals. Its obesity trial, published in JAMA, showed 16.65 percent weight loss.

The 16.6 percent topline was old news. The full SYNCHRONIZE-1 data, out at the diabetes meeting and in the New England Journal of Medicine, showed visceral fat down 34 percent, liver fat down 63 percent, and lean mass largely spared. The glucagon arm is why.

Petrelintide's full Phase 2 data landed at the diabetes meeting. The weight loss trailed the incretins, but vomiting came in below the placebo arm, which is the whole case for the amylin class.

In 13,204 matched pairs of adults with obesity and an autoimmune disease, GLP-1 use was tied to about 31 percent fewer pulmonary embolisms, fewer deep-vein clots and strokes, and nearly half the death rate. Heart attacks and coronary stents, the arterial outcomes, did not move, so the benefit clustered where these patients are already primed to clot.

Christian Heinis's lab built a library of 15,360 random cyclic peptides, made the whole set small and greasy enough to pass a cell membrane, then screened for the rare crossers before asking what they bound. The lead, peptide 30 at 890.6 daltons, blocked the intracellular Keap1-Nrf2 interaction inside living cells. The membrane is the wall that keeps most peptide drugs as injections aimed at surface receptors, and a reliable way through it changes which targets are worth a peptide program at all.

TRANSCEND-T2D-1, published in The Lancet and presented at ADA on Saturday, showed retatrutide cutting A1C by up to 2.0 points and pushing as many as 46 percent of adults with type 2 diabetes back to a normal, non-diabetic blood-sugar level. The same dataset logged seven arrhythmias and three major cardiac events across 403 drug-treated patients versus none on placebo. The counts are too small to settle anything, but retatrutide's glucagon-receptor arm is the obvious mechanism, and the cardiovascular outcomes trial does not read out until 2027.