The ESR1 Mutations: From Bedside to Bench to Bedside

Estrogen receptor (ER)-positive breast cancer accounts for the majority of breast cancer cases and the majority of breast cancer mortality. Despite the efficacy of endocrine treatment, resistance remains a major clinical challenge. As we are increasingly sequencing endocrine-resistant tumors, we are uncovering multiple genomic aberrations enriched after endocrine treatment that are likely drivers of endocrine resistance. These include alterations in the MAPK signaling pathway, such as loss-of-function mutations in NF1 and gain-of-function mutations in ERBB2, and alterations related to transcription, consisting of MYC amplifications and hotspot mutations of CTCF and FOXA1 (1). Although the frequencies of these alterations are increased after endocrine treatment, their prevalence remains relatively low and the most common mutations acquired after endocrine treatment are those affecting the ESR1 ligand–binding domain. In metastatic ER-positive breast cancer, ESR1 mutations are detected in 30%–40% of patients, and of these, the Y537S and D538G mutations are the most common (reviewed in ref. 2). In addition, a number of less prevalent missense hotspot ESR1 mutations were identified. These low-prevalence ESR1 mutations have not been as extensively characterized, but they appear to possess diminished phenotypes compared with Y537S and D538G (3). ESR1 mutations are rare in primary disease and evolve under the selective pressure of aromatase inhibitor treatment in metastatic disease. Although to a lesser extent, a number of studies have now shown that the ESR1 mutations are also enriched after adjuvant endocrine treatment (4, 5).

The discovery of the ESR1 mutations in breast cancer metastatic tumors and the subsequent preclinical and clinical studies underscore the importance of bedside-to-bench-to bedside translational research. The initial preclinical studies of the ESR1 mutations showed the mutations engender ligand-independent transcriptional activity, enhanced growth in hormone-deprived conditions, and relative resistance to tamoxifen and fulvestrant. The structure of the Y537S and D538G mutations was solved and provided mechanistic insights to the constitutive properties of these mutations. In addition, biophysical studies revealed that the mutations have decreased affinity to tamoxifen and fulvestrant, thus explaining their relative resistance to these agents (summarized in ref. 2). In clinical studies, plasma cell-free DNA (cfDNA) analysis of ESR1 mutations in patients with ER-positive metastatic breast cancer demonstrated that these mutations lead to aromatase inhibitor resistance and relative resistance to fulvestrant (6). However, when comparing aromatase inhibitors with fulvestrant in the presence of ESR1 mutations, fulvestrant was shown to have superior activity. Most recently, the initial report of the PADA-1 phase III clinical trial was presented at the 2020 American Society of Clinical Oncology (ASCO) Meeting. The PADA-1 trial, one of the first studies to prospectively study ESR1 mutations in metastatic ER-positive breast cancer, revealed that the presence of ESR1 mutations significantly decreased the median progression-free survival in first-line treatment with an aromatase inhibitor and the CDK4/6 inhibitor, palbociclib [median progression-free survival was 11 months in the presence of an ESR1 mutation and 26.7 months with wild-type (WT) ESR1]. In addition, baseline and end-of-treatment analysis of ESR1 mutations in the PALOMA-3 phase III clinical trial, in which patients with metastatic breast cancer who progressed on an aromatase inhibitor were randomized to palbociclib in combination with fulvestrant versus fulvestrant alone, showed an increase in Y537S mutations after treatment with fulvestrant or fulvestrant in combination with palbociclib. Both studies suggest that the ESR1 mutations have a role in the resistance to the combination of endocrine treatment plus palbociclib. ER acts upstream of cyclin D1–CDK4/6–Rb1 axis, which can explain the role of the activating ESR1 mutations in resistance to endocrine treatment in combination with CDK4/6 inhibition. ESR1 mutations are also a prognostic marker of poor overall survival in metastatic breast cancer. This was demonstrated for both the Y537S and D538G as monoclonal and polyclonal mutations.

A number of groups have generated stable cell lines expressing the ER mutations. Various methods have been employed to develop these models, including models in which the mutations were knocked-in by applying gene editing technologies and homologous recombination and models with stable inducible expression of ESR1 mutations. Despite the differences between these models, these studies all validated the ligand-independent transcriptional activity of the mutations and also showed that the ESR1 mutations regulate the transcription of genes that are not estrogen dependent in ESR1 WT cells (7, 8). In the study by Arnesen and colleagues, the unique ER mutant–regulated genes were enriched in genes involved in cell migration and motility. Similarly, in our study, in which we investigated doxycycline-inducible cell models of the ESR1 mutations, the ER mutant–regulated genes were enriched in genes involved in metastasis. Furthermore, we showed that the expression of the ER mutations promotes metastasis in in vivo studies. Importantly, we showed that silencing of the ER mutations led to regression of metastases, which demonstrates the therapeutic potential of targeting ESR1 mutations (8). The studies from the stable cell lines from the different research groups showed that the unique transcriptomes regulated by ER mutants are allele specific, with functional differences between the Y537S and D538G mutations. The differences between the Y537S and D538G were also borne out by the respective structures, pharmacology, and clinical outcomes of the ESR1 Y537S and D538G mutants.

The study by Arnesen and colleagues (7) also highlights the variability between different cell models developed in different laboratories. As evidenced by the principal components analysis plots of the RNA sequencing of the MCF7 and T47D cell models in the article by Arnesen and colleagues, the cells strongly segregated on the basis of the laboratory of origin and not by mutational status or treatment condition. This is not surprising, as the cells have been maintained in different laboratories and the mutations were introduced by employing different methods. While the doxycycline-inducible models were generated from polyclonal cell populations and enable the evaluation of isogenic models, the models produced by Arnesen and colleagues were derived from single-cell clones. A landmark study demonstrated that cell lines are not clonal or genetically stable (9). Single-cell clones derived from the same MCF7 parental population differed in their mutational landscape, and the clones continued to evolve after multiple passages. Thus, the clonal nature of the cells described in the article (9) may also explain the differences in the transcriptional effects of the mutant ER cistrome observed in this article versus previous work. These differences illustrate the cell context–dependent functional consequences of the ESR1 mutations, which are likely true for other mutations and have important clinical implications. As an example, PIK3CA mutations in early-stage ER-positive breast cancer are associated with good outcomes, whereas in metastatic ER-positive breast cancer, PIK3CA mutations are associated with decreased overall survival compared with WT PIK3CA. Moreover, in metastatic triple-negative breast cancers, PIK3CA mutations are associated with better outcomes (10).

In addition to the cell context–dependent functional consequences of ESR1 mutations, the functional and clinical consequences of ESR1 mutations may vary in the context of different treatment regimens. As aforementioned, there is clinical evidence that ESR1 mutations contribute to and evolve during the onset of resistance to palbociclib and endocrine treatment. However, it is not known whether ESR1 mutations result in unique therapeutic vulnerabilities after the acquisition of resistance to CDK4/6 inhibitors. Along these lines, we do not know whether and how ESR1 mutations affect response to chemotherapy or novel targeted therapies in clinical development, such as AKT1 inhibitors. Importantly, we now have early evidence from the PADA-1 study that successful clearance of the cfDNA ESR1 mutations in first-line treatment of metastatic disease with endocrine treatment and palbociclib was associated with improved outcomes (presented at ASCO 2020). However, it is not known why some patients receiving a certain treatment regimen will clear cfDNA ESR1 mutations, whereas other patients receiving the same treatment do not. This may be due, in part, to the tumor-specific genetic or epigenetic background and/or to the allele frequency of the mutation, or the specific allelic mutation.

Taken together, the preclinical and clinical studies all point to the importance of targeting ESR1 mutations in metastatic ER-positive breast cancer. In response, there are a number of new endocrine treatments in clinical development that will hopefully enable better targeting of ER mutations and improve patient outcomes.

R. Jeselsohn reports grants from Lilly and Pfizer and personal fees from Carrick Therapeutics and Luminex during the conduct of the study. No disclosures were reported by the other author.