Testing ESR1 effects in untreated primary tumors
ESR1 mutations are a known driver of resistance to aromatase inhibitor-based endocrine therapy, and they become common in metastatic ER-positive disease after such treatment. Most mechanistic evidence, however, comes from cell-line models or metastatic cohorts, where prior therapy and metastatic site can confound interpretation. This study instead asked what ESR1 mutations look like in primary, therapy-naïve tumors, before adjuvant treatment could reshape the tumor.
The authors drew on the Swedish SCAN-B population-based cohort. They identified 27 primary breast tumors carrying ESR1 mutations and selected 108 ESR1 wild-type controls matched on ER status, progesterone receptor status, ERBB2 amplification, subtype, and adjuvant endocrine treatment. After quality filtering, the analyzed proteogenomic set comprised 20 mutant and 86 wild-type tumors, quantified across transcriptome, proteome (9857 proteins), and phosphoproteome (8971 phosphosites).
Mutant tumors carry more mutations, not one partner gene
ESR1-mutant tumors had a significantly higher mutational burden than wild-type tumors: the median frequency ratio was 1.92, with a Wilcoxon P below 0.001, and the per-megabase COSMIC mutation count also differed significantly (Wilcoxon P = 4.611 × 10⁻⁴). This was not explained by a single co-mutated driver, since no individual gene was consistently co-mutated with ESR1.
The authors interpret this as a downstream consequence of estrogen signaling, plausibly linked to enhanced proliferation and concurrent DNA replication. Supporting that reading, pathway scores for G2M checkpoint, DNA repair, and unfolded protein response correlated positively with tumor mutational burden in both transcriptomic and proteomic data, while angiogenesis and apical junction scores correlated negatively. Age at diagnosis and Ki-67 levels showed no significant correlation with mutational burden, which argues against a simple clinical explanation.
MTORC1 activation appears in all three data layers
MTORC1 is a kinase complex that links nutrient and growth signals to protein synthesis and cell growth. The study looked for its activity at several levels rather than relying on one measurement. Gene set enrichment comparing mutant with wild-type tumors flagged MTORC1 among pathways enriched in both transcriptome and proteome data.
Phosphoproteomic kinase-substrate enrichment analysis added a functional layer: scores indicated activation of downstream MTORC1 effectors such as RPS6KB1-2, alongside inhibition of upstream inhibitory kinases including AKT1 and PRKACA. Because these are three different readouts, agreement across them is the paper's main argument for genuine pathway activation rather than a transcription-only artifact.
High MTORC1 signaling marks faster relapse and shorter survival
Using median splits of sample-wise pathway scores, the authors ran Cox regression against relapse-free and overall survival. Elevated MTORC1 signaling was associated with shorter relapse-free survival (transcriptome hazard ratio 2.58, P = 0.037; proteome hazard ratio 2.54, P = 0.039) and shorter overall survival (transcriptome hazard ratio 3.11, P = 0.029; proteome hazard ratio 3.07, P = 0.031). Kaplan-Meier curves showed the same direction at both RNA and protein level.
The UV response pathway ran the other way: it was downregulated in ESR1-mutant tumors and associated with better prognosis, though not consistently significant, with an overall-survival hazard ratio of 0.33 (P = 0.033) in transcriptome data and 0.42 (P = 0.083) in proteome data. These are associations within one cohort, not proof that MTORC1 causes the relapses.
ER and MTORC1 programs converge only partially
If mutant ER directly drove MTORC1, ESR1 levels should track MTORC1 activity. They largely did not. Estrogen Response and MTORC1 scores correlated positively within each data layer, but the cross-pathway correlation between Estrogen Response and MTORC1 was only weak to moderate. MTORC1-related kinase activity showed no significant correlation with ESR1 transcript levels, with only weak signals for AKT1 (negative) and PRKCA (positive).
The authors also checked whether mutant tumors simply contain more ER. ER transcript and protein levels correlated strongly with each other but did not differ between mutant and wild-type cases, and ESR1 mutant allele frequency was not significantly associated with ESR1 transcript or ER protein abundance. The paper's interpretation is that MTORC1 activation reflects downstream ER activity and is potentiated by ESR1 mutations in a context-dependent way, rather than being switched on by ER abundance.
A recurrence signature built and tested in one cohort
Combining tumor grade, MTORC1 signaling, and ESR1 mutation status, the authors built a predictive signature that separated ESR1-mutant from wild-type patients and flagged those at high recurrence risk. Feature selection and model training both used the SCAN-B dataset, with leave-one-out cross-validation and DeLong tests comparing models.
That design carries a clear limit: because development and evaluation happened in the same cohort, the reported predictive performance should be read cautiously. The external checks in METABRIC, TCGA, and SweBCG91RT cohorts could not use ESR1 mutation status and instead defined high risk by recurrence within one year, so those results test the signature's behavior under a different definition, not its ability to detect ESR1 mutations.
Where the MTORC1 link matters most
The paper's most decision-relevant point is a comparison, not a single number: MTORC1 signaling, measured at both transcript and protein level, separated outcomes more consistently than ESR1 mutant allele frequency did. Allele frequency showed no significant Cox association with relapse-free survival (hazard ratio 4.61, 95% CI 0.49 to 42.61, P = 0.178) or overall survival (hazard ratio 2.19, 95% CI 0.23 to 21.06, P = 0.497).
So for readers weighing whether ESR1 mutations imply mTOR pathway dependence, the evidence supports a context-dependent potentiation rather than a direct switch, observed in 20 mutant and 86 wild-type untreated primary tumors. The natural next question the data point to is whether MTORC1-high, ESR1-mutant tumors respond differently to pathway-directed therapy, which this cohort cannot answer.
Sources used in this answer
ESR1 mutations are associated with enhanced MTORC1 signaling in a context-dependent manner
In therapy-naïve primary breast tumors, ESR1 mutations were associated with higher mutational burden, multi-layer MTORC1 activation, and worse survival when MTORC1 signaling was high, though MTORC1 activity correlated only weakly with ESR1 transcript levels.