Five science items, from an amatoxin ADC payload SAR study to siderophore-vectored MBL inhibitors, plus three business stories: Genentech–Earendil's AI bispecific pact, Mirum's hepatitis delta Phase 3 win and Tolerance's Tanabe IL-33 in-license.
2026-09-28 · ⚗️Antimicrobial peptides (intracellular mode of action) · Journal of Medicinal Chemistry (ACS)
- Melittin is potently antimicrobial but too haemolytic to be a lead; the authors built 10-, 15- and 20-mer peptides from repeats of its C-terminal WIKRK segment.
- The 15- and 20-mers adopt β-sheet structure in bacterial-mimetic vesicles, with negligible haemolysis and cytotoxicity and strong activity against bacteria and mycobacteria.
- They were active in an E. coli mouse infection model by intravenous dosing and inhibited transcription and translation, suggesting an intracellular mechanism.
💡 PointsShrinking melittin to a tandem repeat and flipping it from helix to sheet is a neat way to decouple killing from lysis. The intracellular mode of action is the interesting part, since it points away from the membrane-disruption liabilities that kill most AMPs in the clinic. I'd still want protease stability and plasma binding data before believing the in vivo window.
β-Sheet Antimicrobial and Antimycobacterial Peptides with a High Therapeutic Index by Employing the Repeats of a 5-Mer Segment of α-Helical Cytolytic Melittin ↗
2026-09-27 · ⚗️α-Amanitin payload (RNA polymerase II) / ADC · Journal of Medicinal Chemistry (ACS)
- The team probed the δ-position of the 4,5-dihydroxyisoleucine residue in α-amanitin, a residue long assumed to drive RNA polymerase II inhibition.
- Two δ-modified analogs carry handles for direct linker attachment, so conjugation no longer has to go through other positions on the bicyclic peptide.
- Both were converted into trastuzumab conjugates that showed extremely high cytotoxicity on target-expressing cells, supporting these amatoxins as ADC payloads.
💡 PointsThis is the kind of payload SAR I want more of. Amanitin's DHIle has been treated as untouchable, so finding a δ-position that tolerates substitution and doubles as a conjugation handle opens a cleaner attachment strategy than the usual lysine or Asn shuffles. I'd want to see plasma stability and bystander behaviour next, since transcription-arrest payloads live or die on therapeutic index.
Unlocking Potent Amatoxin Antibody−Drug Conjugates Using Structure−Activity Studies on δ-Substituted-γ-Hydroxy Isoleucine ↗
2026-09-27 · ⚗️Integrin αvβ6 radiotheranostic peptide · Journal of Medicinal Chemistry (ACS)
- Existing αvβ6-targeting tracers for pancreatic cancer suffer from modest tumor uptake and fast washout; the authors varied hydrophilic linkers and an albumin-binding moiety, guided by docking and MD.
- The lead, NOTA-Asp2-IPB-αvβ6L, was labeled with 18F, 68Ga and 177Lu and kept high hydrophilicity, target affinity and albumin binding.
- In PET/CT it showed higher tumor uptake and longer retention while keeping favorable pharmacokinetics, supporting a matched imaging–therapy pair.
💡 PointsTumor washout is what sinks most small-peptide radioligands, and an albumin binder is the standard fix, but the balance is delicate: too much binding and the blood pool swamps your contrast. Pairing it with a very polar Asp2 linker looks like a sensible way to buy retention without paying in background. The real test is kidney and marrow dose with 177Lu.
Albumin-Binding Modification Improves Tumor Retention of Hydrophilic αvβ6-Targeting Radiotheranostic Tracers ↗
2026-09-27 · ⚗️Metallo-β-lactamase (NDM) inhibitor / siderophore conjugate · Journal of the American Chemical Society (ACS)
- Tris-catecholate siderophore–meropenem conjugates overcame porin-loss and efflux-based carbapenem resistance in P. aeruginosa in an iron-dependent manner.
- The team also vectorized an indole-carboxylate MBL inhibitor, which alone barely works in P. aeruginosa, apparently for permeability reasons.
- The conjugates restored inhibitor activity and reversed meropenem and cefiderocol resistance in MBL-producing clinical isolates, in vitro and in a Galleria mellonella model.
💡 PointsPermeability, not target affinity, is what kills most Gram-negative programs, so vectorizing a known but impermeable MBL inhibitor is a smart shortcut. Reversing cefiderocol resistance is the headline for me, since that drug already exploits iron uptake. I'd press on conjugate stability in serum and whether resistance simply migrates to siderophore receptors.
Siderophore-Enabled Delivery of Carbapenem and Metallo-β-Lactamase Inhibitor: A Promising Strategy against Multidrug-Resistant Pseudomonas aeruginosa ↗
2026-09-28 · 🏢trial result · Fierce Biotech · Mirum Pharmaceuticals
- The Phase 3 portion of AZURE-1 met its primary endpoint at week 24: roughly half of treated patients responded on either the weekly or four-weekly subcutaneous regimen, versus none on delayed treatment.
- No grade 3 or higher or serious adverse events were reported through week 24. A second Phase 3, AZURE-4, is expected in Q4, with a filing possible in the first half of next year.
💡 PointsMirum now has a credible, differentiated challenger in hepatitis delta, with a monthly-dosing option that Gilead's and Vir's programs will have to answer commercially. But the depth-of-suppression numbers are modest, so the AZURE-4 readout matters more than today's headline. For an acquired asset, a clean first Phase 3 also validates the deal thesis and de-risks Mirum's rare-liver pipeline story.
Mirum posts phase 3 hepatitis win for 'commercially different… ↗
2026-09-28 · 🏢licensing deal · Fierce Biotech · Tolerance Bio / Tanabe Pharma
- Tolerance Bio takes global development rights and ex-Japan commercial rights to Tanabe Pharma's clinical-stage anti-IL-33 antibody MT-2990, renamed TLB-33, in a deal with milestones that could reach $560 million.
- The antibody has been tested in more than 150 patients across five trials; Tolerance plans a Phase 2 study on preserving thymic function and immune resilience.
💡 PointsSecond thymus-themed in-license in a month: Tolerance is assembling a pipeline around one biological thesis using de-risked clinical assets rather than its own discovery. That is capital-efficient, but it means the company's value rests on the thymus hypothesis being right, not on any single molecule. Pharma sellers get a clean exit for a deprioritized antibody, the usual pattern for mid-stage assets.
Tolerance continues thymus-focused deals with $560M pact for Tanabe's… ↗
2026-09-24 · 🏢partnership · Fierce Biotech · Genentech (Roche) / Earendil Labs
- Genentech will pay Earendil Labs $55 million upfront, with up to about $1.445 billion in development, regulatory and sales milestones, plus tiered royalties.
- Earendil leads antibody discovery and early clinical work on pre-agreed target pairs using its AI antibody-generation platform; Genentech takes over global development and commercialization.
💡 PointsRoche is buying optionality on target combinations, not a single asset: a small upfront against a milestone-heavy tail is the classic platform structure, and it limits the downside if the AI-generated molecules disappoint. The signal is that big pharma still wants bispecific capacity in oncology after the PD-1×VEGF rush, and is willing to source it from AI-native shops.
Roche's Genentech inks $1.5B bispecific pact with fast-rising Earendil… ↗