ESR1 mutations and therapeutic resistance in metastatic breast cancer: progress and remaining challenges

Abstract

Breast cancer accounts for 25% of the cancers in women worldwide. The most common subtype of breast cancer diagnosed is hormone receptor positive, which expresses the oestrogen receptor (ER). Targeting of the ER with endocrine therapy (ET) is the current standard of care for ER-positive (ER+) breast cancer, reducing the mortality by up to 40%. Resistance to ET, however, remains a major issue for ER + breast cancer, leading to recurrence and metastasis. One major driver of ET resistance is mutations in the ER gene (ESR1) leading to constitutive transcriptional activity and reduced ET sensitivity. These mutations are particularly detrimental in metastatic breast cancer (MBC) as they are present in as high as 36% of the patients. This review summarises the pre-clinical characterisation of ESR1 mutations and their association with clinical outcomes in MBC and primary disease. The clinically approved and investigational therapeutic options for ESR1 mutant breast cancer and the current clinical trials evaluating ESR1 mutations and ET resistance are also discussed. Finally, this review addresses pre-clinical models and multi-‘omics’ approaches for developing the next generation of therapeutics for ESR1 mutant and ET-resistant breast cancer.

Introduction

Breast cancer is one of the most common cancers worldwide, accounting for 25% of the cancers in women and 12% of the cancers in men and women [1]. Over 2 million women were diagnosed with breast cancer worldwide in 2018 [1]. The most common breast cancer subtype is hormone receptor positive, expressing the ER and/or progesterone receptor and accounting for as many as 80% of breast cancer cases [2]. For those women with ER + breast cancer who do not express epidermal growth factor receptor 2 (HER2), targeting ER with ET is the currently recommended strategy [3]. The ET options for postmenopausal ER + primary breast cancer include aromatase inhibitors (AIs) that block the conversion of androgen to oestrogen, or selective oestrogen receptor modulators (SERMs) that antagonize ER activation by reducing co-factor binding [3]. These ETs have been used as adjuvant therapy for up to 10 years, with the SERM tamoxifen (Tam) used for 2–5 years, followed by AI therapy for up to 5 years or an additional 5 years of Tam for patients intolerant to AIs [3].

While the use of ET in primary breast cancer reduces the mortality by up to 40%, breast cancer often becomes resistant to treatment, resulting in breast cancer recurrence and metastasis [4]. Treatment of postmenopausal ER + MBC includes ET in combination with inhibitors for therapeutic targets, such as cyclin-dependent kinase 4/6 (CDK4/6), mammalian target of rapamycin (mTOR), or phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA). The preferred first-line ETs include AIs or the selective oestrogen receptor degrader (SERD) fulvestrant (Fulv) [3]. Multiple therapeutic combinations are available, but therapeutic resistance and eventual death from MBC remain a serious issue. Therefore, it is imperative that we understand the mechanisms of therapeutic resistance that lead to disease progression and ultimately death in MBC. One mechanism of therapeutic resistance in ER + MBC is the acquisition of somatic mutations in ESR1, which are present in up to 36% of MBC patients [5]. Since their discovery in 1997, ESR1 missense mutations have been extensively reported in breast cancer samples, with the majority of them being in the ligand-binding domain (LBD) [6]. Mutations in ESR1 confer constitutive activity to the ER and exhibit decreased ET sensitivity, suggesting that these mutations are a critical player in ET resistance and disease progression for a significant portion of MBC patients.

Pre-clinical characterisation of ESR1 mutations and their role in therapeutic resistance

ESR1 mutations confer constitutive ER activity

ESR1 mutations S47T, K531E and Y537N were initially discovered in MBC samples by Zhang et al. in 1997 [7]. Lentiviral transfection of ESR1 LBD mutation Y537N into breast cancer cell lines conferred oestrogen-independent transcriptional activity that was unaffected by treatment with either Tam or Fulv, indicating an intrinsic resistance to ET [7]. This study hypothesised that the position of the Y537N mutation in the LBD of the ER may induce a conformational change conferring ER ligand independence. The hypothesis that LBD mutations induce ER conformational changes was validated in a study using molecular dynamics simulations of LBD mutations Y537S and D538G [8]. These two mutations enhance the hydrogen bonding of the LBD, resulting in increased stability of the agonist conformation of ER [8]. In addition, cell lines transfected with Y537S or D538G mutations exhibited constitutive transcriptional activity compared to wild-type (WT) ESR1 cells even in the presence of Tam or Fulv, and required significantly higher doses of the drugs to reduce ER activity compared to that of WT cells [8]. An in silico study revealed that mutations in the LBD region of ESR1 adopt an apo conformation even when bound to ER antagonists and SERDs [9]. Further in silico characterisation of ESR1 mutations confirmed that the increased flexibility of mutant ERs reduces ligand-binding affinity and disrupts the bonding patterns to Fulv, which are the mechanisms of the ET resistance [10].

ESR1 mutations also exhibited enhanced recruitment of co-activator proteins in an oestrogen-independent manner, leading to constitutive transcriptional activity [11, 12]. The Y537S mutation enhanced the recruitment of steroid receptor co-activator 3 (SRC3) as a mechanism of increased transcriptional activity [11]. SI-1, an SRC inhibitor, exhibited synergism with the experimental SERD AZD9496 and reduced cell viability in MCF-7 cells expressing the Y537S mutation [11]. SI-2, an SRC inhibitor, in combination with AZD9496 inhibited the tumour growth in a patient-derived xenograft (PDX) model containing the Y537S mutation [11]. MCF-7 Y537S cells developed using CRISPR-Cas9 genome-editing technology have also exhibited oestrogen-independent recruitment of co-activators AIB1 and p300, and elevation of RNA polymerase II and histone H3 acetylation associated with ligand-independent transcription [12]. The reduced ET drug affinity and enhanced recruitment of multiple transcriptional co-activators as a result of the conformational changes induced by ESR1 mutations make treatment of ESR1 mutant breast cancer particularly challenging.

ESR1 mutations exhibit proliferative, stem-cell, epithelial-to-mesenchymal transition (EMT) and metastatic phenotypes

The ligand-independent activity of ESR1 mutant breast cancer manifests as a variety of phenotypes, including enhanced mitogenic signalling from growth factor receptors, stem-cell biology, epithelial-to-mesenchymal transition, and ultimately metastasis. Overexpression cell line models of ESR1 Y537N, Y537S, and D538G mutations exhibited increased insulin-like growth factor 1 (IGF1) signalling. Y537S cells exhibited enhanced binding of the IGF1 receptor (IGF1-R) with ER, and this enhanced binding was associated with resistance to Tam. Interestingly, cells were re-sensitised to Tam with small interfering RNA knockdown of two regulators of phosphoinositide 3-kinase (PI3K), PIK3R1 and PIK3R3, indicating that PI3K pathway alteration is an intrinsic mechanism of ESR1 mutant breast cancer ET resistance [13]. In another study, ESR1 mutant cell lines exhibited an increased IGF1 transcriptional signature and were again reliant on the PI3K pathway, though it was cell line dependent [14]. Importantly, there was no enhanced IGF1 response in Tam-resistant WT cell lines, indicating that IGF1 may be a unique ET escape mechanism for ESR1 mutant cells [14]. While IGF1-R inhibition was synergistic with Fulv in multiple ESR1 mutant cell lines, IGF1-R inhibition has been clinically unsuccessful in MBC and is therefore not yet a viable clinical target [14–16]. Recent in vitro work on MCF-7 cells overexpressing Y537S or D538G ER identified activation of the PI3K–AKT–mTOR pathway using microarray analyses and also increased PI3K and mTOR phosphorylation [17]. The PI3K–AKT–mTOR pathway regulates cell metabolism; MCF-7 Y537S cells exhibited increased mitochondrial activity compared to WT cells and utilised glutamine to promote cellular migration [17]. While IGF1-R inhibition is not a currently viable therapeutic, each of the above studies identified a PI3K pathway reliance. Therefore, targeting the PI3K–AKT–mTOR pathway directly or through downstream metabolic pathways may be an option to exploit this therapeutic vulnerability.

Studies have also linked ESR1 mutations with enhanced stem cell and EMT phenotypes, both of which are associated with metastasis [18, 19]. Cell lines overexpressing ESR1 mutations exhibited enrichment in CD44⁺/CD24⁻ cells compared to WT cells, indicating an increase in the breast cancer stem-cell phenotype [18]. This stem-cell phenotype was associated with increased NOTCH4 expression [18]. Blocking phosphorylation of serine 118 (S118) in ESR1 via genetic mutation of the amino acid reduced NOTCH4 expression and mammosphere formation, indicating that phosphorylation of S118 is a mechanism for promotion of the stem-cell phenotype in ESR1 mutant cells [18]. A recent study has established that Y537S MCF-7 cell RNA expression correlated with EMT and PAM50 basal-like signatures, which have been associated with stem-cell phenotypes [19]. EMT and PAM50 basal-like signatures are associated with shorter recurrence-free survival and poor pathological complete response to neoadjuvant chemotherapy in breast cancer patients [19–21]. This study also utilised in vivo cell-mixing experiments comparing the tumour growth and metastatic frequency associated with different proportions of ESR1 Y537S MCF-7 cells mixed with WT cells [19]. Increased proportions of Y537S mutant cells resulted in enhanced tumour growth and metastatic frequency [19]. In addition, all distant metastases were predominantly mutant cells [19]. An in silico analysis comparing 216 whole-exome-sequenced MBC tumours to The Cancer Genome Atlas primary tumour database identified ESR1 mutations not only as enriched in metastases compared to primary tumours, but also as drivers of metastasis, as all ESR1 mutations were identified in ET-resistant MBC [22].

ESR1 mutations exhibit intrinsic ET resistance and are acquired with ET

A study analysing the adaptation to AI therapy identified ligand-independent ER activity in recurrent tumours treated with AIs, including increased expression of ER-regulated genes [23]. In addition, cell lines overexpressing mutant ER exhibited constitutive transcriptional activity and reduced sensitivity to both Tam and Fulv [24]. A recent work analysing the chromatin immunoprecipitation sequencing profiles of ESR1 mutant cell lines identified unique binding site profiles compared to WT cells [25]. These unique binding site profiles were enriched for Tam resistance using gene set enrichment analysis [25]. The mechanisms of ET resistance in ESR1 mutant cells, therefore, include unique ER-binding profiles in addition to the mechanisms discussed above. A recent work has identified that ESR1 mutations are also acquired in multiple cell lines after long-term oestrogen deprivation. Oestrogen-deprived ESR1 mutant cell lines exhibited ligand independence, altered cistromes and ET resistance phenotypes of overexpression and genetically modified ESR1 mutant models [26]. The reduced sensitivity to ETs and detection of ESR1 mutations after oestrogen deprivation suggest these mutations as key players in ET resistance.

ESR1 mutations and MBC

ESR1 mutations are acquired and enriched in MBC

Given the extensive pre-clinical profiling of ESR1 mutations and their intrinsic ET resistance, many ET-treated cohorts have been screened for ESR1 mutations (Table 1). NGS identified ESR1 mutations that occurred preferentially in metastatic patients, with ESR1 mutations in 12% of metastatic tumours (9/76) versus 0% (0/58) of primary tumours [27]. In another cohort, NGS identified 6/11 patients with ESR1 mutations, all of them after ET [28]. Another study identified enrichment of ESR1 mutations in MBC ranging from 3.5% (11/313) in primary tumours to 13.6% (84/616) in MBC samples [8]. A recent study analysed whole-exome sequencing data from a large collection of ER + HER2− biopsies and identified ESR1 mutations as one of 32 genes significantly enriched in metastases (762 MBC, 739 primary) [29]. ESR1 mutations in circulating tumour DNA also predicted the disease progression of patients treated with AIs better than the measurement of cancer antigens or total levels of cell-free DNA [30]. ESR1 mutations were detected in 31.4% (22/70) of MBC patients in this cohort, and there was a 4.9-fold increase in the risk of progression at 3 months for ESR1 mutant patients, confirming that ESR1 mutations can identify patients at a higher risk for ET resistance [30].

Table 1.

ESR1 mutations identified in metastatic breast cancer patient cohorts.

Publication	Cohort sample information	Patients with ESR1 mutations	Mutation identification method	ESR1 mutations identified
Jeselsohn et al. [27]	ER + /HER2− MBC tumours	9/76 (12%)	Targeted NGS	E380Q, Y537N, Y537C, Y537S, D538G
Robinson et al. [28]	ER + MBC tumours	6/11 (54%)	Whole-exome NGS	L536Q, Y537S, Y537C, Y537N, D538G
Toy et al. [8]	HR + MBC tumours	84/616 (13.6%)	Targeted NGS (MSK-IMPACT assay)	Y537S, Y537N, Y537D, Y537C, V534E, V478L, V418E, S463P, S432L,S329Y, N532K, L536R, L536P, L536H, L466Q, G442R, G344D, F461V, E542G, E380Q, D538G, A546D
Schiavon et al. [5]	Advanced breast cancer treated with AI for the first time in the metastatic setting	16/44 (36.4%)	Multiplex ddPCR	Mutations tested: L536R, Y537C, Y537N, Y537S, D538G
Clatot et al. [30]	MBC treated with first-line AI serial blood draws	22/70 (31.4%)	ddPCR	D538G, Y537S, Y537N, Y537C, E380Q, S463P, L536R
Razavi et al. [29]	HR + /HER2− advanced breast cancer tumours after ET	128/692 (18%)	Targeted NGS (MSK-IMPACT assay)	A546D, D538G, E380Q, E542G, F461V, G160D, G344D, G442R, H524L, K252N, L466Q, L536H, L536P, L536Q, L536R, L540Q, M421V, N407I, N532K, S329Y, S432L, S463P, V138E, V478L, V534E, Y537C, Y537H, Y537N, Y537S
Clatot et al. [31]	Circulating DNA from ER + MBC with disease progression on AIs	44/144 (30.6%)	ddPCR	Y537S, D538G, Y537N, Y537C
Franken et al. [32]	Circulating tumour cells from ER + luminal subtype MBC patients on oestrogen deprivation therapy	13/46 (28%)	Single-cell whole-genome sequencing	G160C, K252R, N348S, E380Q, D426E, L507F, Y537C, Y537N, D538G, D541Q
Zundelevich et al. [33]	MBC or recurrent breast cancer tumour samples	5/41 (12%) newly diagnosed MBC 5/28 (18%) advanced MBC 15/41 (36%) local recurrence	Targeted sequencing or ddPCR	D538G, Y537C, Y537S
Jeannot et al. [34]	Cell-free DNA from HR + /HER2− MBC with disease progression on AIs	11/42 (26%)	Multiplex ddPCR	E380Q, L536H, D538G, Y537N, Y537S, Y537C
Turner et al. [37]	Circulating tumour DNA from HR + MBC with disease progression on a nonsteroidal AI	115/383 (30%)	Multiplex ddPCR	D538G, Y537S, Y537N, Y537C, L536R, S463P, E380Q
Spoerke et al. [55]	Circulating tumour DNA from ER + MBC	57/153 (37%)	ddPCR	E380Q, S463P, P535H, L536Q, L536R, L536H, L536P, Y537C, Y537N, Y537S, D538G
Turner et al. [38]	Circulating tumour DNA from advanced breast cancer progressing on previous therapy	222/1051 (21%)	ddPCR and Targeted NGS	Y537S, D538G, E380Q
Jardim et al. [96]	19,545 breast tumour samples	11%	NGS	Not specified
O’Leary et al. [47]	ER + /HER2- MBC with disease progression on a previous ET treated with fulvestrant with or without palbociclib	49/195 (25.1%) pre-treatment 61/195 (31.3%) post-treatment	ddPCR	Q75E, E380Q, S463P, L536H, L536I, L536P, L536R, L536V, Y537C, Y537N, Y537S, D538G, T553S
Lok et al. [54]	Circulating tumour DNA from ER + /HER2− MBC patients treated with tamoxifen and BCL-2 inhibitor venetoclax	10/33 (30%)	ddPCR	D538G, S463P, I514V, Y537C, E380Q, Y537S
Moore et al. [97]	ER + MBC	89/338 (26%)	Targeted NGS	D538G, Y537S, Y537N, Y537C, E380Q, L469V, L536H, L536P
Razavi et al. [58]	ER + /HER2− MBC tumour tissue	6/38 (15%)	Targeted NGS	Y537S, D538G
Chandarlapaty et al. [60]	Cell-free DNA from ER + MBC treated with AIs	156/541 (28.8%)	ddPCR	D538G, Y537S

ER+ oestrogen receptor positive, HER2− human epidermal growth factor receptor 2 negative, NGS next-generation sequencing, ESR1 oestrogen receptor 1, MBC metastatic breast cancer, AI aromatase inhibitor, MSK-IMPACT Memorial Sloan Kettering-Integrated Mutations Profiling of Actionable Cancer Targets, ddPCR digital droplet polymerase chain reaction, LBD ligand-binding domain, HR+ hormone receptor positive, ET endocrine therapy, BCL-2 B-cell lymphoma 2.

ESR1 mutations are associated with AI therapy resistance in MBC

The selection of ET in MBC is particularly important, as several studies indicate that ESR1 mutations are acquired with AI therapy in MBC. In one study, patients treated with adjuvant AI therapy had detectable ESR1 mutations in 3.8% of the samples, but this number rose to 38% in patients treated with AIs in the metastatic setting [5]. ESR1 mutations were also associated with decreased progression-free survival (PFS) in patients treated with AIs, consistent with pre-clinical data that ESR1 mutations were associated with ET resistance [5]. Another study determined that 30% of the MBC patients treated with an AI had at least one ESR1 mutation at the time of progression, and PFS was lower for ESR1 mutant patients treated with AIs compared to WT patients [31]. Interestingly, ESR1 mutations were not predictive of the outcomes for subsequent ETs [31]. A study using single-cell genomic analysis of circulating tumour cells in plasma samples from MBC identified ESR1 mutations in 28% of the patients, and mutations were only observed in individuals receiving oestrogen deprivation therapy (AIs in postmenopausal women and gonadotropin-releasing hormone agonists in pre-menopausal women) [32]. In another study, ESR1 mutant patients again exhibited reduced PFS with AI treatment and ESR1 mutations were present in up to 18% of the MBC patients [33]. Patients with an ESR1 mutant allele frequency greater than 1% in recurrent breast cancer also exhibited a reduced distant recurrence-free survival [33]. These studies suggest that not only do AIs enhance ESR1 mutant acquisition in MBC, but ESR1 mutant patients also exhibit worse outcomes with AI therapy.

The use of fulvestrant in ER+MBC

While MBC patients acquire ESR1 mutations with greater frequency when treated with AIs and these ESR1 mutations are associated with decreased PFS, the role of ESR1 mutations in therapeutic resistance to SERDs is less clear. As discussed above, pre-clinical characterisation of ESR1 mutations indicates that they require significantly higher concentrations of SERDs than WT cells to reduce ER activity [8]. A study investigating ESR1 mutations in AI-resistant MBC identified no difference in PFS between mutant and WT patients after second-line treatment with Fulv plus the CDK4/6 inhibitor palbociclib [34]. While pre-treatment ESR1 mutation status did not predict the PFS for Fulv plus palbociclib therapy, detection of ESR1 mutations in circulating tumour DNA at 30 days of treatment predicted decreased PFS for Fulv plus CDK4/6 inhibitor palbociclib [34]. Therefore, while baseline ESR1 mutation status may not predict the Fulv therapy response, identifying patients who continue to harbour ESR1 mutations during therapy may be important for identifying those at risk of earlier disease progression.

Large clinical trials, however, have identified minimal differences in PFS for Fulv treatment in ESR1 mutant versus WT MBC patients. The EFECT and SoFEA trials analysed the effect of nonsteroidal AI exemestane versus Fulv on disease progression in ER  +  MBC patients on an AI [35, 36]. A pooled analysis of ESR1 mutant patients (115/383) from these two trials identified a shorter PFS for mutant versus WT in the exemestane arm (2.4 months versus 4.8 months) [37]. Strikingly, there was no difference between ESR1 mutant and ESR1 WT patients in the Fulv arms (3.9 months versus 4.1 months) [37]. This meta-analysis supports previous literature that Fulv exhibits no differences in PFS between ESR1 mutant and WT patients, in contrast to the reduced PFS of mutant patients treated with AIs [31].

The recent plasmaMATCH trial sequenced circulating tumour DNA samples of patients with advanced breast cancer who had progressed on adjuvant ET or at least one line of therapy for advanced breast cancer [38]. Eighty-four ESR1 mutant MBC patients were enrolled on an “extended-dose” regimen of Fulv by doubling the frequency of administration, hypothesising that ESR1 mutant patients are more sensitive to Fulv than alternative ETs and may benefit from an increased dose [38]. The change in tumour size was similar between ESR1 mutant and WT patients [38]. Disappointingly, however, there was no improvement in response rate with dose increase or increased frequency of dosing, with disease progression in 72/76 patients and a median PFS of 2.2 months [38]. The authors do note that their study population was heavily treated and that this may contribute to ET resistance [38]. While many studies identified no difference in PFS between ESR1 mutant and WT MBC patients, treatment resistance to Fulv leading to disease progression remains a major issue for MBC, indicating additional ETs and effective combinatorial treatments are an essential clinical need regardless of mutational status.

The use of CDK4/6 inhibitors with ET for ER + MBC

To improve the therapeutic response, the use of a CDK4/6 inhibitor in combination with first- or second-line ET is now recommended for MBC patients [3]. The current recommended first-line therapy for MBC is an AI in combination with one of three CDK4/6 inhibitors: ribociclib (MONALEESA-2) [39], palbociclib (PALOMA-2) [40] or abemaciclib (MONARCH-3) [3, 41]. Each of these CDK4/6 inhibitors has been associated with improved PFS over an AI alone in women with ER+ advanced breast cancer [39–41]. The use of these CDK4/6 inhibitors with Fulv is also more effective than Fulv alone in advanced breast cancer. The MONALEESA-3 trial investigated ribociclib plus Fulv as a first-line therapy for MBC, and identified a nearly 8-month improvement in PFS [42]. The MONARCH-2 [43] (Fulv versus Fulv plus abemaciclib) and PALOMA-3 [44, 45] (Fulv versus Fulv plus palbociclib) studies recruited patients with relapse or progression on a previous ET for MBC. MONARCH-2 identified that median PFS was extended nearly 6 months (16.4 vs 9.3) with the addition of abemaciclib [43]. Follow-up analyses of the MONARCH-2 cohort showed that abemaciclib in combination with Fulv improved the overall survival by 9.4 months [46]. The addition of abemaciclib also improved the patient quality of life by delaying chemotherapy compared to Fulv alone (50.2 versus 22.1 months) and improved chemotherapy-free survival by 7.3 months [46]. While abemaciclib in combination with Fulv is effective for ET-resistant patients, more work is needed to determine how ESR1 mutant patients may respond to this combination therapy.

Extensive work analysing ESR1 mutations has been done for the combination of palbociclib and Fulv. Initial analyses of the PALOMA-3 cohort identified an improved median PFS of 5.4 months for Fulv plus palbociclib compared to Fulv alone [44]. Follow-up analyses of patients in PALOMA-3 did, however, identify an enrichment of Y537S ESR1 mutations at the end of therapy from 25.1 to 31.3% (both treatment arms combined) [47]. ESR1 mutant patients also exhibited a reduction in PFS of 6.3 months compared to ESR1 WT patients in the Fulv plus palbociclib arm [47]. The PFS improvement with combination therapy versus Fulv alone, however, is similar between mutant and WT patients (mutant: hazard ratio = 0.43 versus WT: hazard ratio = 0.49) [48]. In addition, a follow-up analysis of early circulating tumour DNA from the PALOMA-3 cohort indicated that ESR1 mutations were not predictive of PFS for CDK4/6 inhibitor treatment [49]. Therefore, while treatment with palbociclib plus Fulv shows worse outcomes in ESR1 mutant MBC patients compared to WT patients, the improved efficacy over Fulv alone indicates that palbociclib combination therapy does improve patient outcomes. Clinical trials are underway (discussed below) to evaluate the use of other CDK4/6 inhibitors in ESR1 mutant MBC, particularly abemaciclib.

Additional targeted therapies for the treatment of ER + MBC

Several targeted therapy candidates for MBC treatment have been investigated for the emergence of ESR1 mutations and therapeutic efficacy. Both JQ1 (bromodomain and extra-terminal motif [BET] inhibitor) and vorinostat (a class I and II histone deacetylase inhibitor) [50] enhanced tumour regression in combination with Fulv, though not as monotherapies, in a D538G ESR1 mutant xenograft model developed from circulating tumour cells [50]. JQ1 plus Fulv also reduced the metastatic burden in lungs, though primary tumour growth reduction was similar to other combinations [50]. Unfortunately, multiple trials using BET inhibitors in solid tumours have been terminated due to the variable pharmacokinetics and limited efficacy (NCT02698176 [51] and NCT02431260 [52]). A small study analysing heavily treated MBC evaluated vorinostat in combination with an AI and identified the general treatment tolerability and clinical benefit of either treatment response or stable disease in 6/16 patients [53]. Though the study size was small, the clinical benefit of HDAC inhibitors for heavily treated MBC patients requires further investigation (NCT01153672) [53]. With a reduction in both metastases and tumour growth in pre-clinical ESR1 mutant models, chromatin-modifier inhibitors with improved toxicity profiles may be highly effective for the treatment of ER + MBC.

MBC patients heavily pre-treated with AIs were also responsive to modulation of the apoptotic response using the B-cell lymphoma 2 (BCL-2) inhibitor venetoclax [54]. Venetoclax, though ineffective alone, enhanced response to Tam with an objective response rate of 45% [54]. Seventy percent of the ESR1 mutant patients had a tumour response or prolonged stable disease, and a corresponding reduction in ESR1 mutations in circulating tumour DNA after treatment with venetoclax and Tam [54]. Expanded trials using inhibitors of this apoptotic target are necessary for ESR1 mutant patients to determine if it may be an effective combination.

A recent work in ESR1 mutant MBC has also identified the co-occurrence of ESR1 mutations with PIK3CA mutations, suggesting PI3K inhibitors as a possible therapeutic route for a subset of ESR1 mutant patients. The FERGI trial identified 14.1% more ESR1 mutations in PIK3CA mutant patients compared to PIK3CA WT patients [55]. Another study identified a co-occurrence of ESR1 mutations in 24% of PIK3CA mutant tumours [56]. Although the small-molecule PIK3CA inhibitor taselisib exhibited pre-clinical efficacy in ESR1 mutant models [56], it was halted as a prospective therapeutic due to variable toxicity and only a 2-month improvement in PFS versus Fulv alone in the SANDPIPER clinical trial [57]. A recent analysis of the PI3K inhibitor alpelisib in combination with AI therapy identified that MBC patients with ESR1 mutations showed a lack of clinical benefit (P = 0.0067) [58]. Pre-clinical evaluation of alpelisib and oestrogen deprivation revealed that ESR1 mutant cell line xenograft tumours were significantly less responsive to therapy than WT [58]. The authors do note, however, that the association of AI resistance with ESR1 mutant MBC suggests that AI resistance rather than alpelisib resistance may be responsible for the disease progression of the ESR1 mutant patients in this study [58]. Evaluation of other ETs such as Fulv with alpelisib in ESR1 mutant MBC is necessary to determine if ESR1 mutations also exhibit intrinsic resistance to PI3K inhibitors.

Another targeted therapy to combat AI therapy resistance is the mTOR inhibitor everolimus. The TRINITI-1 PhaseI/II trial evaluated the dose-limiting toxicity and PFS associated with the CDK4/6 inhibitor ribociclib in combination with the AIs exemestane and everolimus [59]. Patients with locally advanced or MBC that had progressed on one to three lines of ET, but no more than one line of chemotherapy were eligible for the trial [59]. Analysis of circulating tumour DNA in these patients identified a reduced PFS with the triplet therapy for patients with ESR1 mutations [59]. A secondary analysis of the BOLERO-2 trial evaluating everolimus plus the AI exemestane in MBC patients also determined that patients with ESR1 mutations had a reduced overall survival compared to WT patients regardless of the treatment [60]. Importantly, both the BOLERO-2 and TRINITI-1 trials evaluated everolimus in combination with an AI, which may be the factor contributing to the reduced therapeutic efficacy in ESR1 mutant patients. In addition, a subset analysis of patients in the BOLERO-2 trial harbouring only the D538G mutation showed a similar PFS for WT patients in the everolimus combination arm [60]. Therefore, there is evidence that targeting the mTOR pathway in combination with ET may improve the PFS of MBC for a subset of ESR1 mutant patients.

Regardless of the targeted therapy chosen, the underlying intrinsic resistance to ET remains a challenging roadblock for treating ESR1 mutant MBC. While sequential use of ETs may enhance the survival of patients, eventual progression is inevitable. Therefore, the development of additional novel ETs such as oral SERDs with pre-clinical efficacy in ESR1 mutant models is of paramount importance. In addition, detection of the ESR1 mutation is associated with decreased efficacy of ET, and it is therefore important to understand the mechanisms associated with resistance in ESR1 mutant patients to identify additional clinically relevant targets for combination therapies to improve the outcomes. As sequencing technologies advance, further characterisation of ESR1 MBC through mutli-‘omic’ approaches may provide additional insight into the mechanisms of ET resistance.

ESR1 mutations and primary breast cancer

While ESR1 mutations are clearly enriched and contribute to ET resistance in MBC, a recent study assessed whether ESR1 mutations in primary breast cancer could serve as a predictive biomarker of ET response in early breast cancer [61]. Though ESR1 mutations were rare in the SCAN-B primary breast cancer cohort (30/3217 ER+ patients), they were associated with worse relapse-free survival (P = 0.011) and overall survival (P = 0.019) [61]. ESR1 mutations also predicted poor relapse-free survival and overall survival for patients receiving ET (P = 0.007, P = 0.010, respectively) [61]. Another study using whole-exome sequencing of biopsy samples from ET-resistant patients (809 ET-naive, 642 post-treatment) revealed that, of all the mutations identified, ESR1 mutations had the greatest enrichment in post-ET-treated tumours [29]. ESR1 mutations were also associated with decreased PFS for AI-treated but not SERD-treated patients [29]. Most interestingly, ESR1 mutations were present in 18% of ET-resistant samples and were mutually exclusive at the genomic level with mutations in the mitogen-activated protein kinase pathway, the proto-oncogene Myc, or additional transcription factors [29]. The high frequency and mutual exclusivity of ESR1 mutations implies that they are a genomically distinct and significant subgroup driving ET resistance in primary breast cancer. These studies support that even though ESR1 mutations are rare in primary breast cancer, screening for them may be important for the identification of early breast cancer patients likely to develop ET resistance and relapse.

Clinical trials for the evaluation of ESR1 mutations in MBC and potential new therapeutics

Clinical trials evaluating ESR1 mutations in patients treated with AIs+/− CDK4/6 inhibitors

The FMER trial identified ESR1 mutations in 31.4% (22/70) of the MBC patients with progressive disease after first-line AI treatment and a 4.9-fold increase in early progression during AI treatment [30]. Several additional trials are active or recruiting participants to confirm that ESR1 mutant acquisition contributes to therapy resistance with AI alone or in combination with targeted therapies (Table 2). The CICLADES trial is currently active and monitoring the circulating tumour DNA of several common mutations including ESR1 as a predictive biomarker for early detection of ET resistance in MBC patients treated with AIs (NCT03318263 [62]). Another trial is analysing tissue and blood samples to predict response to AIs with palbociclib as a first-line therapy for MBC (NCT03439735 [63]). The outcomes include time to ESR1 mutant acquisition with AI plus palbociclib, and evaluation of ESR1 mutant status before, during, and after therapy [63]. Another study is recruiting a cohort of Asian women with MBC with a confirmed Y537C mutation, and will analyse their outcomes compared to WT, D538G, and Y537S ESR1 mutant patients on any ET (NCT04212702 [64]). While these trials are important to identify additional patients with resistance to AIs due to ESR1 mutant acquisition, clinical trials are especially needed to evaluate the role of ESR1 mutations in second- and third-line therapies after progression on AIs.

Table 2.

Clinical trials evaluating ESR1 mutations with an aromatase inhibitor and fulvestrant-based therapies.

	Clinical trial Identifier	Trial name	Status	Trial arm(s)	Patient attributes	Primary outcome	Mutation testing	Sample collection timepoints
AI-based Therapy	NCT03318263	CIrCuLAting Dna ESr1 Gene Mutations Analysis (CICLADES)	Active, not recruiting	AI treatment	MBC	ESR1 mutation incidence	Yes: NGS exon sequencing of circulating tumour DNA. Mutations: ESR1 any PIK3CA any AKT any	(1) Baseline (2) Three months before progression (3) Time of progression
	NCT03439735	Determinants of Resistance to First-line Therapy with an AI and Palbociclib for HR + MBC	Recruiting	AI + palbociclib	HR + MBC	ESR1 mutation measure prior to therapy	Yes: sequencing of plasma tumour DNA. Mutations: ESR1 any	(1) Baseline (2) Timepoints throughout treatment
	NCT04212702	ESR1 mutations in Asian ER + Metastatic Breast Cancer on Hormonal Therapy-based Treatments	Recruiting	Homornal therapy-based treatments	ER + MBC	Percentage of patients with ESR1 LBD mutations	Yes: NGS targeted sequencing of ESR1 Mutations: ESR1 Y537C, Y537S, D538G	Baseline
Fulv-based therapy	NCT03079011	PAlbociclib and Circulating Tumour DNA for ESR1 Mutation Detection (PADA-1)	Active, not recruiting	(A) AI + palbociclib (B) Fulv + palbociclib (C) Fulv + palbociclib AFTER progression on Arm A (AI + palbociclib)	ER + /HER2- MBC treated with first-line AI + palbociclib	(1) Safety (2) PFS for change in therapy in patients with rising ESR1 mutations treated with AI + palbociclib	Yes: sequencing of circulating tumour DNA Mutations: ESR1, any	Regular intervals during treatment
	NCT02738866	Palbociclib With Fulvestrant for Metastatic Breast Cancer After Treatment With Palbociclib and an Aromatase Inhibitor	Recruiting	Fulv + palbociclib	HR+/HER2− MBC	(1) PFS with new therapy after progression with AI + palbociclib 2) ESR1 and PI3K mutation prevalence	Yes: sequencing of tumour biopsies and plasma tumour DNA. Mutations: ESR1, any PI3K, any	Not specified
	NCT03560856	A Biomarker Study of Palbociclib + Fulvestrant for Second, and Third Line of Postmenopausal Women With HR + /HER2- Advanced Breast Cancer (PALPETBIO)	Not yet recruiting	Fulv + palbociclib as second- or third-line therapy	HR+/HER2− advanced breast cancer	PFS with therapy	Yes, unspecified method Mutations: ESR1, any Additional Biomarkers—unspecified	Not specified
	NCT02913430	Phase II Treatment of Metastatic Breast Cancer With Fulvestrant Plus Palbociclib or Tamoxifen Plus Palbociclib	Recruiting	(A) Fulv+ palbociclib (B) tamoxifen + palbociclib	ER + MBC with 2–3 prior lines of ET	PFS with therapy	Yes, sequencing of tumour biopsy and plasma samples Mutations: ESR1, any	Prior to and following treatment (up to 5 years)

AI aromatase inhibitor, Fulv fulvestrant, MBC metastatic breast cancer, LBD ligand-binding domain, HR+ hormone receptor positive, ER+ oestrogen receptor positive, HER2− human epidermal growth factor receptor 2 negative, ESR1 oestrogen receptor gene, PIK3CA phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha, AKT protein kinase B, PI3K phosphoinositide 3-kinase, NGS next-generation sequencing, PFS progression-free survival.

Clinical trials evaluating CDK4/6 inhibitors with fulvestrant in ESR1 mutant MBC

Given the intrinsic resistance of ESR1 mutations to ET and the increased efficacy of adding palbociclib to Fulv in MBC as a second-line therapy for ESR1 mutant MBC patients (PALOMA-3) [48], clinical trials are currently evaluating the ESR1 mutant frequency and outcomes with this combination (Table 2). The PADA-1 trial is evaluating ESR1 mutations in circulating tumour DNA from patients treated with palbociclib in combination with an AI as a first-line MBC therapy (NCT03079011 [65]). Patients with enrichment of ESR1 mutations will be randomised to continue AI plus palbociclib, change to Fulv plus palbociclib, or progress on AI plus palbociclib with follow-up treatment of Fulv plus palbociclib. Another trial is evaluating the acquisition of ESR1 and PIK3CA mutations in patients treated with Fulv plus palbociclib after progression on an AI plus palbociclib (NCT02738866 [66]). ESR1 mutations are acquired, enriched and develop resistance to AIs, so these studies are invaluable to determine if CDK4/6 inhibitors should be continued with additional lines of ET. Although ESR1 mutations were acquired in both treatment arms of PALOMA-3, the combination of palbociclib with Fulv still improved the PFS for ESR1 mutant patients compared to those on Fulv alone [47]. It will be enlightening to see if the results of PADA-1 indicate that ESR1 mutant patients exhibit improved PFS with Fulv compared to AIs in combination with palbociclib, or if there is clinical benefit to continuing palbociclib treatment in combination with Fulv after progression with palbociclib plus AIs. A confirmatory study of the PALOMA-3 trial is also planned to determine if genomic alterations or a genomic signature can identify patients likely to develop resistance to Fulv plus palbociclib (NCT03560856 [67]). Finally, another trial is evaluating MBC patients with 2–3 prior lines of ET and their responses to Tam plus palbociclib versus Fulv plus palbociclib (NCT02913430 [68]). While patient treatment will not be pre-selected based on ESR1 mutation status, they will be measuring ESR1 mutations in circulating tumour DNA throughout the treatment. The results of these trials will be paramount to guide the selection of ET combinations, particularly for heavily pre-treated ESR1 mutant patients.

Novel SERMs/SERDs and targeted therapies for ER

Patients with MBC currently receive 2–3 lines of ET before combinations with chemotherapy are recommended [3]. For patients with a higher risk of ET resistance, such as AI resistance in ESR1 mutant MBC, additional ET options are needed. Evaluation of several novel SERMs and SERDs is underway in clinical trials (Table 3). The SERM lasofoxifene was found to be efficacious in vitro and in vivo in ESR1 mutant models, including reduction of metastases [69]. A clinical trial is currently recruiting participants to evaluate PFS with lasofoxifene versus Fulv in patients with ESR1 mutant MBC who have progressed on an AI plus CDK4/6 inhibitor (NCT03781063) [69]. The ELAINEII trial will be evaluating lasofoxifene plus abemaciclib in patients with ESR1 mutant MBC after progression on first- or second-line ET (NCT04432454 [70]). These two trials are not only important for the evaluation of a SERM that is efficacious specifically for ESR1 mutant patients, but ELAINEII is also one of the first trials to evaluate abemaciclib specifically in ESR1 mutant MBC patients, as most current trials are investigating palbociclib. Another novel SERM, elacestrant, will be compared to standard-of-care ET in the EMERALD trial recruiting ER+/HER2− MBC patients who have progressed on 1–2 lines of ET, including an AI or Fulv plus a CDK4/6 inhibitor (NCT03778931 [71]). The EMERALD trial will analyse the PFS in ESR1 mutant versus WT patients [71]. Finally, the SERM bazedoxifene plus palbociclib is being assessed in ER + MBC (NCT02448771) [72]. Preliminary results identified the combination as well-tolerated, and genomic profiling to evaluate ESR1 mutations is underway [73]. Given the evidence for AI resistance in ESR1 mutant MBC, the option of additional SERMs as first-line or second-line ET could significantly expand ET options for ESR1 mutant MBC patients.

Table 3.

Novel endocrine and ER-targeted therapies in development.

Therapy name	Type of therapy	ESR1 mutant pre-clinical efficacy?	Clinical trials	Clinical trial arm(s) (if applicable)
Lasofoxifene	SERM	Yes, unspecified mutations	(1) NCT03781063 (2) NCT04432454	(1) Lasofoxifene versus fulvestrant in ER + /HER2− MBC with an ESR1 mutation Recruiting (2) Lasofoxifene + abemaciclib in ER + /HER2− MBC with an ESR1 mutation Recruiting
Elacestrant	SERM	Y537S and D538G [91]	NCT03778931	Elacestrant versus standard of care ET in ER + /HER2− advanced breast cancer after CDK4/6 inhibitor therapy Active, not recruiting
Bazedoxifene	SERM	Y537S and D538G [92]	NCT02448771	Bazedoxifene + palbociclib in advanced HR + /HER2− breast cancer Active, not recruiting
AZD9833	SERD	In vivo D538G [74]	(1) NCT04214288 (2) NCT04711252	(1) Three different doses of AZD9833 versus fulvestrant in ER + /HER2- metastatic or recurrent breast cancer Recruiting (2) A comparative study of AZD9833 plus palbociclib versus anastrozole plus palbociclib in patients with ER-positive HER2 negative breast cancer who have not received any systemic treatment for advanced disease (SERENA-4) Recruiting
GDC-9545	SERD	In vivo Y537S [76]	NCT04546009	GDC-9545 + palbociclib versus letrozole + palbociclib in ER + /HER2− MBC or locally advanced breast cancer Recruiting
SAR439859	SERD	In vivo Y537S [93]	(1) NCT04059484 (2) NCT04478266	(1) SAR439859 versus endocrine monotherapy of choice in ER + /HER2− locally advanced or MBC Recruiting (2) Amcenestrant (SAR439859) plus palbociclib as first-line therapy for patients with ER ( + ) HER2(−) advanced breast cancer (AMEERA-5) Recruiting
LY3484356	SERD	In vivo Y537S [94]	NCT04188548	LY3484356 + /− abemaciclib, everolimus, or alpelisib. Triplet combinations of LY348456 + abemaciclib + AI or LY348456 + abemaciclib + trastuzumab in advanced or metastatic breast cancer or endometrial cancer (EMBER) Recruiting
ZN-c5	SERD	In vivo Y537S [95]	(1) NCT04176757 (2) NCT03560531 (3) NCT04514159	(1) Zn-c5 as a monotherapy in ER + HER2− breast cancer Recruiting (2) ZN-c5 + /− palbociclib in ER + HER2− advanced breast cancer Recruiting (3) ZN-c5 + /− abemaciclib in ER + HER2− advanced breast cancer Recruiting
37D	B-SERD	In vitro Y537S D538G [80]	n/a	n/a
ER-148	ER degrader small molecule	In vitro Y537S D538G [81]	n/a	n/a
BHPI	ER biomodulator	In vitro Y537S D538G [82]	n/a	n/a

ESR1 oestrogen receptor gene, ER+ oestrogen receptor positive, HER2− human epidermal growth factor receptor 2 negative, MBC metastatic breast cancer, SERM selective oestrogen receptor modulator, SERD selective oestrogen receptor degrader.

Planned clinical trials testing novel therapies targeting the ER and pre-clinical candidates exhibiting in vitro efficacy.

While evidence suggests that ESR1 mutant tumours are more responsive to Fulv than other ETs, the lack of oral bioavailability of Fulv necessitates intramuscular injections, which are an increased burden on patients. Several novel oral SERDs are currently in clinical trials and pre-clinical development (Table 3). A new oral SERD, namely AZD9833, phenocopied the pre-clinical efficacy of Fulv both in vitro and in vivo, including tumour reduction in a PDX harbouring a D538G mutation [74]. Therefore, AZD9833 may be an efficacious SERD option for ESR1 mutant patients. While there are no ESR1 mutant-specific clinical trials for AZD9833, the SERENA-2 trial is currently recruiting patients with advanced ER+/HER2− MBC to evaluate the PFS of AZD9833 versus Fulv (NCT04214288) [75]. The novel SERD GDC-9545 exhibited anti-proliferative activity in both ESR1 mutant and WT cell lines and anti-tumour activity in PDX tumours [76]. GDC-9545 is being evaluated in combination with palbociclib for improved PFS versus the AI letrozole plus palbociclib in patients with ER+/HER2− MBC (NCT04546009) [77]. A small cohort of patients with ESR1 mutations were given the oral SERD SAR439859 with acceptable toxicity and promising anti-tumour activity [78]. The Phase 2 AMEERA-3 trial will study ESR1 mutant acquisition and PFS for SAR439859 versus standard of care ET in ER+/HER2− MBC (NCT04059484 [79]). While most oral SERD clinical trials are in recruiting stages, the evidence for enhanced efficacy in ESR1 mutant pre-clinical models is positive news for increasing the ET options to treat ESR1 mutant MBC.

Several novel SERMs, SERDs and ER-targeted therapies are not yet in clinical trials but show potential in pre-clinical models (Table 3). A novel class of SERDs with the addition of a basic amino acid sidearm (B-SERDs) decreased the 3D spheroid culture viability of both Y537S and D538G cell lines and were found to be as effective as Fulv in reducing in vivo tumour growth in an ET-resistant mouse xenograft model [80]. Importantly, B-SERDs exhibited brain bioavailability in vivo in contrast to Fulv, identifying a role for these novel B-SERDs in patients with brain metastases [80]. An orally bioavailable ER degrader small molecule, ER-148, exhibited ER downregulation comparable to Fulv, and downregulation of ER-regulated genes in Y537S and D538G cell lines [81]. Though ER-148 currently only exhibits in vitro efficacy, its oral bioavailability and efficacy in ESR1 mutant cell lines make it a promising class of drugs to refine for clinical use. Another unique ER biomodulator, BHPI, targets both the unfolded protein response/protein synthesis and inhibitors of cell death regulation [82]. Treatment with BHPI hyperactivated the unfolded protein response, resulting in blockage of > 90% of protein synthesis and cell death in both D538G and Y537S mutant cell lines, identifying it as a potent ER-targeted therapy for ESR1 mutant cells [82]. Though clinical usage of these ER-targeted therapies is limited, their efficacy in ESR1 mutant pre-clinical models warrants further investigation regarding their potential as alternative therapeutics for patients resistant to current ETs.

Pre-clinical modelling of ESR1 mutant breast cancer: identifying the next round of therapeutics

While ESR1 mutations have been extensively characterised in cell lines, tumours are much more heterogenous. The development of PDX and patient-derived organoids (PDOs) from patient tissue has been a great advancement of the last decade in modelling breast cancer, including its heterogeneity. Multigenerational PDXs conserved phenotypes from their tumours of origin, including DNA copy number and breast cancer subtype derived from gene expression analysis, and metastatic potential [83]. Development of ER + PDX is a challenge, however, with one early study resulting in no ER + PDX [84]. A larger cohort of tumours showed successful engraftment for 13% (7/54) of the ER+/HER2− tumours [85]. Importantly, this cohort of PDX were from the advanced disease category and included ESR1 mutations, amplifications and fusions [85]. The development of additional ER + PDX is essential to improve pre-clinical modelling of ER + breast cancer.

The advent of single-cell sequencing technologies is also critical for studying clinical and pre-clinical ER + breast cancer. Single-cell sequencing has been used for tracing both genomic and transcriptomic evolution of tumours, which is critically important for identifying the mechanisms of therapy resistance. The algorithms developed to analyse single-cell genomic data can infer how cell populations expand and contract in a tumour to construct evolutionary “tumour trees” [86]. Single-cell DNA and RNA-sequencing of pre- and post-treatment neoadjuvant chemotherapy breast cancer samples revealed the outgrowth of select chemo-resistant cells and identified that chemotherapy-resistance transcriptional patterns were acquired during therapy rather than pre-existing [87]. Single-cell RNA-sequencing also revealed therapeutic vulnerabilities in breast cancer metastases developed from PDX models. Three triple-negative breast cancer PDXs were analysed using single-cell RNA-sequencing and revealed dysregulation of oxidative phosphorylation in lung metastases [88]. Blocking oxidative phosphorylation reduced metastases, but did not affect primary tumour size or engraftment [88]. In addition, there is evidence that breast cancer tumours do contain polyclonal ESR1 mutations [5, 30, 60, 89]. Single-cell sequencing will be critical to identify and transcriptionally characterise these polyclonal tumours, as different ESR1 mutations can vary in their therapeutic response [60] and molecular characteristics [7, 8]. Application of these technologies to ER + tumours may be ideal to simulate both ET resistance and metastatic development in pre-clinical models.

While PDX recapitulates the genomic and transcriptional phenotypes of the parental tumours and are a valuable model to study the metastatic development and efficacy of drugs, in vivo models require large amounts of time and resources. Development of ex vivo PDOs from patient samples offers shorter timelines for evaluating the pre-clinical efficacy of the growing list of novel ETs and targeted therapies before testing them in vivo. A study of 95 PDOs determined that they can survive more than 20 passages, conserve the morphology, histology, and genomic profiles of patient samples, and developed successfully regardless of histological subtype, in contrast to PDX [90]. Tam treatment response of PDOs developed from MBC samples also reflected patient response, though these results are limited due to the small sample size [90]. Each patient-derived pre-clinical model has strengths and limitations in recapitulating clinical outcomes. Combinations of these models, such as therapeutic screening in PDOs followed by efficacy studies in PDX, will significantly advance the ability to model ER + breast cancer, including therapy resistance.

Conclusion

Since the discovery of ESR1 mutations in MBC in 1997, improved pre-clinical modelling and genomic analysis of extensive collections of patient samples have heavily implicated these mutations in ET resistance. There is a clear acquisition of and selection for ESR1 mutations in AI-treated MBC. The role of SERDs in ESR1 mutant MBC remains unclear, though Y537S mutations were enriched in patients treated with Fulv. The use of CDK4/6 inhibitors in combination with ET in the treatment of ESR1 mutant MBC is a critical area for further study. Extensive work is being done to develop novel SERDs and identify targeted therapies that are efficacious in ESR1 mutant pre-clinical models to improve the outcomes of ESR1 mutant MBC patients, though an “Achilles heel” for these mutations remains elusive (Fig. 1).

Fig. 1. Potential therapeutic targets identified in clinical and pre-clinical studies for ESR1 mutant breast cancer.

Therapeutics are identified in red. Targeting of the PI3K pathway can be achieved through upstream IGF1-R inhibition, direct PI3K inhibition, or downstream inhibition of mTOR or metabolic functions. The ER biomodulator BHPI both inhibits ER and stimulates the UPR to facilitate cell death. The small-molecular inhibitor ER-148 stimulates ubiquitination of ER via recruitment of E3-ligase. The ER co-activator SRC3 can be disrupted by the SRC inhibitor SI-2. Modulation of the cell cycle is an effective strategy that acts by inducing cell cycle arrest by inhibition of CDK4/6. Epigenetic modification of acetylation may be an effective strategy to inhibit HDACs or BET proteins. In addition, clinical trials are currently evaluating the efficacy of oral SERDs and next-generation SERMs to expand endocrine therapy options in ESR1 mutant breast cancer. ER oestrogen receptor, SERD selective oestrogen receptor degrader, Ub ubiquitin, NG SERMs next-generation selective oestrogen receptor modulators, IGF1-R insulin-like growth factor 1 receptor, CDK cyclin-dependent kinase, UPR unfolded protein response, SRC3 steroid receptor co-activator 3, CoA co-activator, HDAC histone deacetylase, BET bromodomain and extra-terminal motif protein, PI3K phosphoinositide 3-kinase, AKT protein kinase B, mTOR mammalian target of rapamycin, BCL-2 B-cell lymphoma 2, Ac acetylation.

Though ESR1 mutations are present in a low frequency in primary tumours, new evidence suggests that they are still predictive of shorter PFS and recurrence. The ESR1 mutation status, therefore, may be relevant for identifying patients at risk of developing ET resistance in both metastatic and primary breast cancer. The disparity of ESR1 mutant frequency between primary tumours and MBC does invite the question of what makes these mutations so prevalent at metastatic sites compared to the breast. Perhaps the stem-cell and EMT phenotypes make ESR1 mutant cells ideal candidates to migrate to metastatic sites and grow compared to other cells in the primary tumour. Most importantly, since these ESR1 mutations are so strongly associated with MBC and ET resistance, what are their therapeutic vulnerabilities? Genomic profiling of patients has associated ESR1 mutations with ET resistance and disease progression, but has not yet elucidated the targetable mechanisms to enhance the treatment of these patients. Initiatives to collect multi-‘omics’ data from patients and pre-clinical models, such as RNA and protein expression in addition to mutational profiles, may be crucial to define the mechanisms of ET resistance in ESR1 mutant breast cancer. An enhanced picture of the resistance mechanisms in ESR1 mutant MBC will be essential to expand the therapeutic arsenal and improve the outcomes of these patients.

Author contributions

Concept and design of review: SKH and SAWF. Writing and review of the paper: SKH and SAWF.

Funding information

This study is supported by BCRF 19-055, NIH R01 CA207270, NIH R01 CA072038, CPRIT MIRA RP180712 (Kelly Hunt, PI), and T32 CA203690-02 to SAWF. SKH received training support from the National Institute of Health (NIH) Ruth L. Kirschstein National Research Service Award (NRSA) Individual Predoctoral Fellowship (F31CA260983) and the NIH Training Program in Cell and Molecular Biology (T32GM008231).

Data availability

Not applicable.

Competing interests

SAWF is a subject editor, member of the Editorial Board of BJC and guest editor of the special issue on metastasis.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.