Background:

ESR1 mutations and fusions arise in hormone receptor positive (ER+ and/or PR+) breast cancer (HR+ BC) patients after aromatase inhibitor (AI) therapy (low estrogen states), to become constitutively active in a ligand-independent manner. Patients with tumors harboring ESR1 D538G/Y537 mutations (by digital droplet PCR of ctDNA) exhibit worse prognosis and outcome with no particular predictive role in determining benefit of chemotherapy vs tamoxifen treatment after progression on AI (Augusto, et al. ASCO 2016). A retrospective analysis of the ESR1 mutation frequency and co-occurring alterations that could guide subsequent therapy approaches was investigated.

Methods:

Molecular profiles of 416 breast tumors [HR+ (n=237), HER2+ (n=29) and TNBC (n=139)] were assessed.Protein expression (IHC) and gene amplification (ISH) were performed.Genomic testing included 592-gene hybrid-capture NGS [NextSeqIllumina platforms] and ArcherDxfusion assay based on anchored multiplex PCR (AMP) FusionPlexSolid Tumor.ESR1 variant (ESR1var) (mutation/fusion) profiles and HR+ BC patients lacking genomic ESR1 alterations (ESR1 WT) were compared; Pearson’s chi-squared test was used to test for significant differences.Results: An ESR1 mutation (point mutations, insertion-deletions [n=49] and fusions [n=4]) was detected in 13% (53 /416) of the specimens, and this constitutes 21% (50/237) of all HR+ breast cancers. Two TNBC patients exhibited ESR1 variants (H398Y in exon 7 and A491S in exon 9, both are classified as variants of unknown significance). ESR1 mutations were not detected in HER2+ BC. Seventy-seven percent of patients with ESR1 mutations were detected in metastatic specimens (p=0.03), with liver (19/53 or 36%) and bone (8/53 or 15%) specimens as the most frequent sources for ESR1 variant (ESR1var) positivity. The most common alleles detected were: D538G (24%), Y537S (18%), E380Q (14%) and L536H (4%); 15 other additional alleles were detected (each 2%). ESR1 fusions were detected in 4 ESR1 WT patients: ESR1-ATP2B2, ESR1-MKL1/ESR1-TNRC6B, ESR1-ARNT2 and ESR1-C6ORRF211. We next compared ESR1var (mutation/fusion) profiles to HR+ breast patients lacking genomic ESR1 alterations (ESR1 WT). ER expression was present in 100% and 96% of ESR1var and ESR1 WT BC, respectively, however expression of PR was negative in 22% and 38% of ESR1var and WT BC, respectively (p=0.05). Significantly higher rates of other gene amplification events observed in ESR1var vs. ESR1 WT BC included: c11orf30 [EMSY, BRCA2 interacting transcriptional repressor] (20% vs. 7%), CCND1 (51% vs.28%), CCND2 (6% vs. 0%), FGF3 (37% vs. 16%), FGF4 (40% vs. 12%) and FGF19 (40% vs. 15%), whereas cMYC was more frequently amplified in ESR1 WT BC (17% vs. 2%); all p-values <0.05. KRAS mutations were higher in ESR1var vs WT BC (4% vs. 0%; p=.004). Alterations in the PIK3CA pathway were common in both ESR1var and ESR1 WT BC: mutations in PIK3CA, AKT and PTEN were observed in 27% and 26%; 4% and 7%; 6% and 5%, respectively, and PTEN loss by IHC in 29% and 26%.

Conclusions:

ESR1 mutations and fusions are detected in 21% of HR+ BC, the majority of which are in metastatic sites. Amplifications of genes involved in downstream regulatory pathways were present and may contribute to the poor prognosis of ESR1var HR+ BC. Correlation with antecedent therapy is currently underway.