The authors examined how metformin influences alternative splicing of ESR1 and the abundance of candidate splicing regulators in breast cancer cell models. The work compared parental MCF-7 cells with a tamoxifen-resistant subline, TAMR1. The central analytic approach was RT-PCR of ESR1 spliceforms paired with immunoblot measurements of splicing-related proteins after exposure to metformin across a 0.25–5 mM dose range. The source did not report detailed seeding density values as numeric cell counts in the abstract; it reported results grouped as high or low density instead.
Across the stated dose range at high seeding density, investigators observed no reproducible change in the share of full-length ESR1 transcript in either cell line. Gel densitometry produced the following percentages for full-length ESR1: MCF-7—42%, 43%, 42%, 43%, 45% and 48%; TAMR1—55%, 44%, 52%, 47%, 68% and 43%. Those values varied without a clear dose-dependent trend. The authors therefore reported that the previously described direction of metformin’s effect on ESR1 splicing does not replicate under these high-density culture conditions.
At low seeding density, results diverged. The MCF-7 cells retained a response to metformin out to 48 hours, with full-length ESR1 shares reported as 52%, 62% and 68% at 0, 2.5 and 5 mM respectively. The abstract did not supply additional timepoints or replicate counts for these measurements; those methodological details were not reported in the summary. For TAMR1 cells at low density, the direction of the splicing change described earlier was observed but conflicted with expectations based on measured changes in candidate proteins (see below).
Protein analyses showed that HMGA1a levels fell in MCF-7 cells following metformin exposure. The TAMR1 line showed a similar tendency, at both seeding densities examined. The authors have previously shown that HMGA1a can promote production of ERα46, the N-terminally truncated ESR1 isoform, by trapping U1 snRNP at an upstream pseudo-5′ splice site. By that mechanism, reduced HMGA1a predicts a relative decline in ERα46 (and therefore a relative increase in full-length ERα). In TAMR1 low-density cultures, however, the observed splicing direction did not match that prediction, which raises mechanistic questions.
The investigators found lower RBM3 protein in TAMR1 cells at 24 and 48 hours of metformin exposure. Given RBM3’s prior implication in metformin-mediated splicing changes in other disease contexts (for example, myotonic dystrophy in earlier literature cited by the authors), the study team suggested RBM3 is a more plausible mediator of the ESR1 splicing effect than HMGA1a in these cells. The abstract did not quantify fold-changes or provide statistical metrics for RBM3 depletion; such details were not reported in the summary.
The authors refer to RT-PCR data from xenograft tumors in an earlier study. Those gels showed full-length ESR1 transcript in the untreated and tamoxifen-plus-metformin groups but not in the single-agent groups; amplification failed in several lanes and the panels do not separate treatments. The summary states these panels are inconclusive for treatment-specific patterns. The abstract does not report numbers of tumors analyzed, nor does it provide amplification controls or statistical treatment for those xenograft gels.
Data indicated that the direction and magnitude of metformin’s effect on ESR1 splicing depend on culture density and exposure time, and that the effect seen earlier does not survive a dose-response test at high density. Based on the measured protein changes, the authors judged HMGA1a an unlikely primary mediator for the observed ESR1 splicing alterations, while RBM3 reduction appeared a more credible candidate in TAMR1 cells. The abstract does not list full experimental replication numbers or p-values; those methodological details were not provided in the text available.
The preprint does not report specific numeric seeding densities, full time-course matrices for every dose, nor statistical parameters for the reported gel densitometry and immunoblot changes in the abstract; these omissions limit immediate interpretation. Also unresolved is why HMGA1a reduction would predict less ERα46 yet the TAMR1 low-density data move in the opposite direction—this mechanistic discordance was not resolved in the summary. The xenograft RT-PCR panels were described as failing to separate treatments cleanly and included lanes with amplification failure, which weakens translational inference. Further work should present full methods, replicate counts, and statistical analyses to clarify whether RBM3 or other factors mediate the density-dependent splicing response to metformin.
Limitations: the source abstract lacks detailed methods (exact seeding densities, full time- and dose-matrices, sample sizes, and statistical reporting), and the xenograft data presented were inconclusive in the panels described. The authors declared no competing interests in the preprint.