Command Palette

Search for a command to run...

Preprint WatchMildNovember 6th, 2026

Metformin lowers HMGA1a and RBM3 protein in breast cancer cells, and its effect on ESR1 splicing depends on cell density and exposure time

Harada, Y.; Tanaka, T.; Hamaguchi, Y.; Horita, Y.; Harada, H.; Mori, M.; Murata, Y.; Ohe, K.

The reported metformin-driven shift in ESR1 exon 1 splicing, which sets the ratio of full-length estrogen receptor to the truncated ER46 isoform, is conditional on seeding density and does not survive a dose-response test; HMGA1a is unlikely to be the splicing factor carrying it.

Mild contradiction

1 prior failure

One documented clinical failure (Phase 1 or 2) overlaps with the claimed mechanism.

This preprint reports that a previously described metformin effect on ESR1 exon 1 splicing, which sets the ratio of full-length oestrogen receptor to the truncated ER46 isoform, does not reproduce under a full dose range at high seeding density, and that HMGA1a is unlikely to be the factor carrying it. Claidex holds one ESR1 failure, vepdegestrant-esr1-protac-ribociclib-tactive-u-breast-cancer-phase1-2-strategic-termination, a Phase 1/2 degrader combination in ER-positive HER2-negative advanced breast cancer stopped by strategic reprioritisation. The connection is indirect, since the preprint concerns isoform regulation rather than degrader pharmacology. It is worth recording because ESR1-directed degraders are evaluated against a single receptor pool, and an N-terminally truncated isoform whose abundance shifts with culture conditions complicates both target quantification and preclinical resistance work. Severity is MILD on one prior non-efficacy failure.

Abstract excerpt

The reported metformin-driven shift in ESR1 exon 1 splicing, which sets the ratio of full-length estrogen receptor to the truncated ER46 isoform, is conditional on seeding density and does not survive a dose-response test; HMGA1a is unlikely to be the splicing factor carrying it.

Matching Claidex post-mortems

1 of 1 indexed

This is an automated contradiction flag, not an editorial judgment on the preprint's quality. Flags identify where the preclinical literature and the clinical failure record diverge.