Therapeutic resistance to anti-oestrogen therapy in breast cancer

Abstract

The hormone receptor oestrogen receptor-α (ER) orchestrates physiological mammary gland development, breast carcinogenesis and the progression of breast tumours into lethal, treatment-refractory systemic disease. Selective antagonism of ER signalling has been one of the most successful therapeutic approaches in oncology, benefiting patients as both a cancer preventative measure and a cancer treatment strategy. However, resistance to anti-oestrogen therapy is a major clinical challenge. Over the past decade, we have gained an understanding of how breast cancers evolve under the pressure of anti-oestrogen therapy. This is best depicted by the case of oestrogen-independent mutations in the gene encoding ER (ESR1), which are virtually absent in primary breast cancer but highly prevalent (20–40%) in anti-oestrogen-treated metastatic disease. These and other findings highlight the ‘evolvability’ of ER⁺ breast cancer and the need to understand molecular processes by which this evolution occurs. Recent development and approval of next-generation ER antagonists to target ESR1-mutant breast cancer underscores the clinical importance of this evolvability and sets a new paradigm for the treatment of ER⁺ breast cancers.

Introduction

Expression of oestrogen receptor-α (ER) is a defining feature of hormone-sensing luminal cells in the mammary gland. Upon activation by oestrogen, ER executes a transcriptional programme that drives proliferation and expansion of the mammary epithelia during puberty¹ (Fig. 1a). Although not all cells of the normal breast gland express ER, ~70–80% of tumours that emerge from the mammary gland express ER and harbour some dependence on ER signalling for proliferation^2–4. ER targeting has thus become the cornerstone of treatment in both early-stage and metastatic ER⁺ breast cancers^5,6. Although most ER⁺ tumours initially appear sensitive to anti-oestrogen agents, resistance can develop over time, characterized by clinical relapse, which typically involves the emergence of genetic and epigenetic alterations that allow for reactivation of ER signalling⁷.

Fig. 1 |. Oestrogen receptor-α signalling and modes of inhibition.

Oestrogen receptor-α (ER) signalling can be inhibited via multiple therapeutic strategies. a, ER is a multidomain ligand-inducible transcription factor that when activated drives a transcriptional programme that supports progression through the cell cycle. b, Strategies for ER targeting that are clinically approved are shown on the top. Aromatase inhibitors, which deplete oestrogen via inhibition of CYP19A1, are also known to select for oestrogen-independent ESR1 mutations. Selective ER modulators inhibit the ligand-binding domain (LBD) but not DNA-binding domain (DBD) or activation function-1 (AF1) domain of ER. They can also act as a partial ER agonist in some contexts. Selective ER degraders inhibit both LBD and AF1 domains by preventing recruitment of co-activators, immobilizing and destabilizing ER, leading to degradation. Approaches for inhibiting ER that are currently under investigation are shown on the bottom. Covalent ER antagonists react to cysteine C530 in the ER ligand-binding pocket and enforce an antagonist conformation in both wild-type and mutant ER. ER-specific PROTACs recruit E3 ligase to the ER, promoting poly-ubiquitination and proteasomal degradation. (Note: ER forms a homodimer but is depicted here as a monomer for visual clarity). Ub, ubiquitin.

Currently approved anti-oestrogen therapies use two main strategies: depletion of ER-activating ligands and/or direct modulation of ER itself (Fig. 1b). The aromatase enzyme (CYP19A1) is required for the synthesis of oestrogens in non-ovarian tissues, and aromatase inhibitors can thus be leveraged to suppress ER signalling in post-menopausal women whose ovaries no longer produce circulating hormones⁸. In pre-menopausal women, oestrogen is the product of both CYP19A1-independent synthesis in the ovary and CYP19A1-dependent synthesis peripherally, and thus aromatase inhibitors are given together with inhibitors of ovarian hormone production such as gonadotropin releasing hormone analogues. As an alternative approach to targeting synthesis of the activating ligand, selective oestrogen receptor modulators (SERMs) can inhibit ER transcriptional activity by directly binding to the C-terminal ligand-binding domain (LBD) of ER and preventing the recruitment of co-activators to this domain⁹. However, SERMs still allow for the engagement of ER with chromatin and therefore do not inhibit transcription driven by the N-terminal ER activation function-1 (AF1) domain^10,11. As AF1 activity is sensitive to various cellular inputs, the impact of SERMs on ER transcriptional output depends on the tissue environment¹². For instance, among the most widely utilized SERMs, tamoxifen generally attenuates oestrogen-stimulated ER signalling in the breast, but exerts agonistic or partial agonistic effects on the uterus, bone and heart tissues, which may be beneficial in some settings (for example, dampening of the osteoporosis process) but detrimental in others (for example, increasing the risk of endometrial cancer)^9,13. Importantly, the agonistic potential of tamoxifen can also manifest in breast cancer cells, and this property presents a liability for therapeutic resistance. Specifically, continuous in vivo passaging of breast cancer cells in the presence of tamoxifen generates tumours in which tamoxifen acts to stimulate cell proliferation through activation of ER signalling^14,15. Recognizing that further benefit might be brought to patients by avoiding the potential for ER agonism, pure oestrogen antagonists were sought through screening for oestrogen competitors that lack ER agonist potential in the rodent uterus¹⁶. Fulvestrant was the result of this campaign for pure anti-oestrogens, and its naming reflects its signalling properties as a full oestrogen antagonist¹⁷. Subsequent to its identification, fulvestrant was observed to promote ER degradation via the ubiquitination–proteasome pathway, leading to the compelling hypothesis that full ER antagonism is achieved through ER degradation and to the designation of fulvestrant as the prototypical selective ER downregulator (SERD)^9,18. Subsequently, it has been proposed that ER degradation induced by SERDs may not be critical to its ER inhibitory effect¹⁹. Instead antagonism is achieved by degradation-independent suppression of the co-activator protein binding to the LBD, together with intranuclear ER immobilization, which suppresses activity from the AF1 domain of ER^20–22.

The growth inhibitory effect of anti-oestrogen agents on ER⁺ breast cancers is due to cellular arrest in the G1 phase of the cell cycle. Given that perturbation of ER signalling results in broad transcriptional reprogramming, precisely how ER antagonists achieve G1 arrest is not fully understood. However, it is thought to involve, at least in part, suppression of cell cycle regulators such as Myc and cyclin D1. Cyclin D1 acts with cyclin-dependent kinase 4 (CDK4) and CDK6 to regulate activity of the retinoblastoma tumour suppressor protein (RB1), which in turn controls gene expression through the E2F family of transcription factors that promote G1-to-S cell cycle progression^23–27.

To date, the most important predictive biomarker of response to endocrine therapy is the nuclear expression of ER, and it is unequivocal that ER-targeted therapies have brought tremendous benefit to the ER⁺ patient population^28–31. However, individual patients with that same predictive biomarker appear to derive widely varying benefits from endocrine therapy. For instance, trials of hormonal therapy in ER⁺ metastatic breast cancer have identified subsets of patients with tumours that progress within 3 months, as well as those who can remain on therapy without relapse for many years^32–35. The clinical vignette of this is striking: different patients may have the identical diagnosis of ER⁺ metastatic breast cancer, with similar prognostic and predictive biomarkers, yet the same targeted treatment may enable tumour control for decades in some patients, whereas others immediately progress. To some extent, the degree of ER expression associates with the degree of efficacy of endocrine therapy. For instance, patients with tumours that express a low level of ER (defined as 10–20 fmol ER per mg protein) derived benefit from a 5-year-course of adjuvant tamoxifen (relative risk (RR) of recurrence = 0.67, standard error (SE) = 0.08), when compared to those with tumours that express a negligible amount of ER (defined as <10 fmol ER per mg protein, which roughly corresponds to an ER scoring of 1%) (RR = 0.97, SE = 0.05). However, this benefit is far from proportional and is only slightly increased even in those with tumours expressing the highest amounts of ER (>200 fmol ER per mg protein) (RR = 0.52, SE = 0.07)^28,36. From a modelling perspective, simply expressing ER is not known to drive tumour growth in the manner of other oncogenes³⁷. As such, ER expression is more of a lineage-specific biomarker reflecting a transcriptional state poised to respond to ER activity, with various other modifiers altering the degree of dependence. This clinical heterogeneity implies that, rather than a singular phenomenon, hormonal therapy resistance in ER⁺ breast cancer represents a spectrum of disease states for which deeper biologic characterization is needed to understand and potentially overcome these clinical disparities.

Mechanisms of resistance to anti-oestrogen therapy

Endocrine therapy is used in various contexts in the care of patients with breast cancer. In the early-stage setting, endocrine therapy may be utilized to control tumour growth, reduce tumour size pre-operatively or prevent distant relapses post-operatively. In the advanced setting, endocrine therapy is given to prolong patient survival and palliate symptoms. Biological evidence for cooperative oncogenic signals such as from the PI3K–AKT pathway has led to the development of combinations of anti-oestrogens with other targeted therapies such as PI3K pathway inhibitors, or CDK4/CDK6 inhibitors, that have enhanced the degree and duration of therapeutic efficacy over anti-oestrogen therapy alone. Thus, the definition of what constitutes ‘endocrine resistance’ may vary on the basis of the particular clinical contexts. Moreover, as ER regulates various oncogenic effectors (for example, G1/S-specific cyclin-D1 (CCND1), Myc and so on), and as each effector can be activated by other signalling pathways, anti-oestrogen therapy can also be overcome by a wide variety of mechanisms^38,39. It is therefore of little surprise that ever-increasing mechanisms of resistance to endocrine therapy have been identified (Fig. 2). Still, these may be thought of as either (1) reactivating ER signalling despite drug administration or (2) bypassing ER to drive another oncogenic programme (Table 1). Interestingly, resistance mechanisms relating to active ER signalling are more typically observed in tumours that manifest with a period of apparent sensitivity followed by resistance, that is, resistance that is acquired under the pressure of therapy, whereas ER-independent mechanisms often appear in tumours that are intrinsically resistant. This suggests that tumours that are initially dependent on oestrogen for proliferation commonly evolve to reactivate ER-directed signalling under therapeutic pressure of anti-oestrogens. The continued dependence of these tumours on ER represents a major, and highly tractable, opportunity to bring further benefit to patients through the development of next-generation ER-targeted therapeutics.

Fig. 2 |. Treatment of ER⁺ breast cancer with endocrine therapies creates a strong selective pressure that drives tumour evolution.

Tumours experiencing the pressure of endocrine therapies can evolve via distinct paths and can be categorized by their dependency on oestrogen receptor-α (ER) signalling (described from left to right below). Dark blue: a subset of cancers retain dependency on ER signalling but acquire features that render them resistant to first-line therapies such as aromatase inhibitors (AIs). The acquisition of ESR1 mutations exemplifies this evolution route, allowing tumours to progress in an ER-dependent but oestrogen-independent or other ligand-independent manner. These ER-dependent tumours are responsive to next-generation ER therapeutics, such as oral selective ER downregulators (SERDs). Yellow: tumours can also acquire alterations in other key transcription factors, such as FOXA1 or MYC. It remains unclear whether transcription factor-altered tumours retain ER dependency, thus further study is required to assess whether they might respond to next-generation ER-targeted therapies or require alternative treatments. Pink and red: other cancers lose dependency on ER altogether. Pink: of these, some become driven by orthogonal signalling pathways, such as HER2 or mitogen-activated protein kinases (MAPKs), and may benefit from the corresponding targeted therapies (for example, anti-HER2 therapies). Red: others lose ER expression altogether and show some evidence of undergoing lineage alterations. These tumours are reminiscent of lineage-plastic prostate cancers, which lose androgen receptor expression, although much less is known regarding lineage plasticity in breast cancer. Further investigation is required to devise treatment strategies for this subset. mBC, metastatic breast cancer; RTK, receptor tyrosine kinase.

Table 1 |.

Common molecular mechanisms of endocrine resistance in breast cancer

Route of resistance	Mechanism	Examples	Mutation prevalence
Retain ER signalling in the presence of ER-pathway-targeted endocrine therapies	Acquisition of ligand-independent ESR1 mutations	LBD or activating ESR1 mutations	0–5% pre-ET, 30–40% post-ET
		ESR1 in-frame fusions with partner genes (e.g., YAP1)	0–2%
	Alterations in the ratio of ER-associated co-regulatory proteins	Elevated expression of co-activators, such as NCOA1 and NCOA3	Not applicable
		Loss of co-repressors, such as N-CoR1 and N-CoR2
Activation and acquired dependency of alternative proliferative pathways	Acquired ERBB2 activity and dependency	ERBB2 amplification; occurs before endocrine therapies and confers intrinsic resistance	~5%
		ERBB2 mutations; acquired following endocrine therapy	1–2% pre-ET, 5–10% post-ET
	Acquired activity and dependency on other receptor tyrosine kinases (RTKs) and downstream MAPK signalling	Amplification of EGFR	1–5% per alteration
		Co-amplification of FGFR1 and associated factors
		Loss-of-function mutations in negative regulators of the MAPK pathway (e.g., NF1)
		Activating mutations in positive regulators of the MAPK pathway (e.g., KRAS, BRAF and MAP2K1)
Broader alterations in the landscape of transcriptional regulators	Mutations in other key transcription factors	Amplifications and hotspot mutations in MYC and CTCF	1–10% per alteration
		FOXA1 mutations in the Wing2-region; proposed to enhance ER-mediated transcription
		FOXA1 mutations (e.g., SY242CS), which activate alternative transcriptomes via non-canonical DNA binding
	Decrease or loss of ER expression, possibly owing’ to lineage plasticity (as in prostate cancer)	Abnormal DNA methylation of the ER CpG island	Not applicable
		Chromatin inactivation by histone deacetylation
		Repression of ESR1 transcription via Twist

Major routes of endocrine resistance, along with examples of specific mechanisms. For mechanisms driven by genomic alterations, literature-reported prevalences are shown. (Disparities in publicly available data sets — for example, assay, sampling method and so on — make it difficult to align mutational frequencies across studies; such meta-analyses are beyond the scope of this Review.) EGFR, epidermal growth factor receptor; ER, oestrogen receptor; ERBB2, erb-b2 receptor tyrosine kinase 2; ESR1, oestrogen receptor 1; ET, endocrine therapy; FGFR1, fibroblast growth factor receptor 1; LBD, ligand-binding domain; MAPK, mitogen-activated protein kinase; NCOA1, nuclear receptor co-activator 1 (also known as SRC1); N-CoR, nuclear receptor co-repressor; NF1, neurofibromin 1; YAP1, Yes1-associated transcriptional regulator.

ER-dependent resistance mechanisms

The evolving capacity of ER⁺ breast cancers is best illustrated by those cancers with acquired, activating mutations in ESR1, the gene encoding ER. Although first described as a case report in 1997 (ref. 40), ESR1 mutations were largely neglected in subsequent years, as initial genomic studies were focused on primary tumours before exposure to therapy and failed to capture the importance of this genetic event. It was not until 2013 that we and others sequenced tumours that had been exposed to therapy in the metastatic setting and demonstrated that acquired mutations in the LBD of ESR1 are important drivers of therapeutic resistance^41–45. Across several large-scale analyses, activating ESR1 mutations are rarely found in primary breast tumours (<1%) or treatment-naive metastatic disease (<5%)^45–47, but are present in a substantial number (~20–40%) of previously treated metastatic tumours, particularly those treated with aromatase inhibitors^{45,46,48–50}. These data strongly suggest that ESR1 mutations are predominantly acquired under the selective pressure created with oestrogen deprivation. Indeed, when sequential samples were taken from the same patients with ER⁺ metastatic breast cancer, ESR1 mutations appear to be selected by aromatase inhibitor treatment and persist through subsequent therapies^51–53.

Most ESR1 mutations are located in the LBD of ER, with mutations occurring at residues E380, L536, Y537 and D538 accounting for the vast majority of all variants detected in previously treated ER⁺ breast cancer. Spanning amino acids 304–554, the LBD contains 12 α-helices (named h1–h12) linked together by loop regions. Without ligand, the LBD is bound by heat shock proteins such as HSP90 and therefore kept in an inactive conformation. In the presence of oestrogen, the LBD dimerizes, ER engages with chromatin and the α-helix h12 forms a hydrophobic groove that binds to co-activator proteins to promote the expression of ER target genes. Binding of the SERM tamoxifen to the LBD prevents h12 from adopting the active conformation, thereby inhibiting recruitment of co-activators to this domain⁵⁴. SERDs likewise bind the LBD and induce a conformation that prevents the recruitment of co-activators; these molecular events occur independently of ER degradation²⁰. Recent structural and biochemical assays, as well as molecular dynamics modelling, have demonstrated that ESR1 mutations stabilize ER in an agonist conformation even in the absence of oestrogen⁵⁵. Mutant ER recruits co-activators in a manner that mimics oestrogen-activated ER, enabling oestrogen-independent expression of ER target genes^46,56,57. This cumulates in a growth advantage for cells harbouring ESR1 mutations, specifically in oestrogen-deprived conditions^41–46. In addition, ESR1 mutations have been shown to enhance migration, invasion and metastatic capacity^43,58,59.

Transcriptomic analyses of preclinical models engineered to express ESR1 mutations have revealed that in addition to oestrogen-independent activation of canonical ER target genes, cells with different ESR1 mutations (for example, those encoding ER.E380Q, ER.D538G and ER.Y537S, within the LBD region) demonstrate mutant-specific gene signatures^41,60–63. The paucity of major structural differences in mutant ER compared with oestrogen-stimulated wild-type ER (as described earlier) makes it challenging to understand how mutated ER can alter cell biology beyond oestrogen independence. We speculate that these ascribed ‘neomorphic’ properties might result from the effects of constitutive mutant ER activity, when compared with periodic oestrogen stimulation of wild-type ER, but more research is necessary to fully understand the molecular mechanisms leading to mutant effects beyond oestrogen independence.

As mutant ER possesses oestrogen-independent activity, oestrogen deprivation strategies such as aromatase inhibitors have proven less effective⁶⁴. Importantly, in addition to enabling oestrogen independence, the mutant-induced agonist conformation reduces the binding affinity of ER antagonists. The SERMs tamoxifen, raloxifene, bazedoxifene, as well as the SERD fulvestrant, all demonstrate decreased binding affinity to LBD-mutated ER proteins when compared with wild type (up to >60-fold)^54,55,65,66. Indeed, the different ER variants (for example, ER.D538G and ER.Y537S) exhibit varying degrees of oestrogen-independent ER activity, and this phenotype roughly correlates with the reduction in ER-drug affinity for each of those variants, as both stem from the receptor pre-folding bias^46,54. The predicted consequence of reduced binding affinity for direct ER antagonists on mutant versus wild-type ER (rather than completely ablated drug binding) is that inhibition of mutant ER may still be achieved by these drugs, but that higher drug concentrations would be required to do so, relative to what is required to inhibit wild-type ER. Indeed, preclinical in vitro data demonstrate that higher concentrations of tamoxifen and fulvestrant are needed to achieve the same degree of inhibition of cell proliferation in the context of ESR1 mutations relative to wild-type ESR1 (refs. 41–46). This phenomenon, whereby higher drug concentrations are required to inhibit mutant versus wild-type ER, is likely most clinically meaningful for drugs that exhibit in vivo exposure constraints, meaning that there is an upper limit on the concentration of drug that can be achieved in patients. Specifically, fulvestrant lacks oral bioavailability, which necessitates administration of high volumes of drug substance (10 ml) via intramuscular injection, most commonly into the gluteus maximus. Imaging studies have shown that this dosing regime is not sufficient to achieve saturation of ER binding in all patients⁶⁷. Given that the exposure constraint of fulvestrant has been proposed as a liability even against wild-type ER, the presence of ESR1 mutations probably further exacerbates this limitation⁶⁸.

The totality of data thus suggests that ESR1 mutations are a major source of resistance to aromatase inhibitors through their ability to activate ER signalling in an oestrogen-independent manner, and also presents a resistance liability for fulvestrant and tamoxifen, owing to their impact on drug binding affinity. The response of the research community to these challenges is described in a later section (entitled ‘Overcoming therapeutic resistance’).

In addition to ESR1 point mutations, structural variants have also been reported for ESR1, and these may also be relevant for therapeutic resistance. Specifically, the first six exons of ESR1 occur as in-frame fusions with various partner genes. For example, ESR1 can fuse with the Hippo pathway co-activator YAP1 or the protocadherin 11 X-linked gene (PCDH11X) to produce stable, functional protein products^44,69. These fusion products lack the ER LBD but maintain the N-terminal transactivation domain and the central DNA-binding domain; they appear to function as oestrogen-independent transcription factors that drive an ER-directed proliferative and metastatic phenotype. By virtue of lacking the LBD, to which all-known ER antagonists bind, these fusion variants are insensitive to the SERM tamoxifen and the SERD fulvestrant. However, it has been suggested that inhibitors of CDK4/CDK6 retain activity against cells expressing these ESR1 fusions^44,56,69. Many other fusion partners with ESR1 have been identified, including AKAP12, GYG1, SOX9, MTHFD1L, FLKHG1, TFG, NKAIN2 and CDK13. A major challenge for studies of ESR1 fusions is that most widely utilized exon-capture sequencing platforms do not reliably identify them and even when other sequencing methods (for example, whole genome sequencing) identify them, their functionality is not clear without additional information such as their level of expression. As such, their clinical prevalence as a resistance mechanism is difficult to quantify and remains a topic of active investigation^70–72.

Outside alterations in the ESR1 gene, resistant tumours can augment ER signalling by changing the ratio of co-regulatory proteins. More than just bridging molecules that help ER form large protein complexes and coordinate spatiotemporal interactions, co-regulators often have intrinsic or associated enzymatic activities, such as acetylation, methylation or ubiquitination, that serve to modulate the transcription of ER target genes⁷³. These can be exploited to confer a hormonal therapy resistant phenotype. For instance, among patients receiving tamoxifen, high expression of the co-activators nuclear receptor co-activator 1 (NCOA1, also referred to as steroid receptor co-activator 1 (SRC1)) and NCOA3 (also referred to as SRC3) have been associated with both therapy resistance and worse prognosis^74–76. Conversely, loss of the ER co-repressors nuclear receptor co-repressor 1 (NCoR1) and NCoR2 (also known as SMRT) led to tamoxifen-mediated proliferation rather than inhibition⁷⁷, and the low NCoR1 mRNA expression level is linked to poorer survival in patients treated with tamoxifen⁷⁸.

Alterations in other signalling pathways

Mechanisms of endocrine therapy resistance in ESR1 wild-type tumours are likely highly fragmented across multiple pathways and mechanisms. This perhaps contributes to the major challenges in fully categorizing and understanding these modes of resistance. However, the activation of HER2 (encoded by ERBB2) signalling is one resistance mechanism in ESR1 wild-type tumours that has been deeply characterized and impacts current clinical treatment paradigms. ERBB2 amplification is found in ~13% of ER⁺ breast cancer^2,3 and has a transcriptional signature that reflects, at least in part, activated HER2 signalling⁷⁹. ERBB2-amplified tumours, found in both the primary and the metastatic settings, often exhibit therapeutic resistance to endocrine therapies that is apparent upon initial drug exposure. Thus, ERBB2 amplification is the classic example of intrinsic resistance to ER-directed therapeutics^80,81. Tumours that are ER⁺ and exhibit amplification of ERBB2 are therefore recognized as a distinct breast cancer subtype (ER⁺ HER2⁺), and these patients will receive HER2-directed therapies. This is an excellent example of how understanding a particular mechanism of therapeutic resistance markedly impacts patient care.

In contrast to ERBB2 amplifications that are found in primary and metastatic tumours before endocrine therapy, ERBB2 point mutations are specifically enriched in ER⁺ tumours that have been exposed to endocrine therapy in the metastatic setting. This represents a mechanism of resistance that may be acquired under the selective pressure of anti-oestrogens^49,82. Tumours with ERBB2 mutations show reduced sensitivity to oestrogen deprivation, the SERM tamoxifen, the SERD fulvestrant as well as inhibitors of CDK4/CDK6. Why ERBB2 amplification exists as a mechanism of intrinsic resistance before endocrine therapy exposure, and ERBB2 point mutations are more commonly an acquired mechanism, is not currently clear and will require further research. The distinction between ERBB2 amplification and point mutations is clinically relevant, as not all HER2-directed therapies that are active against ERBB2-amplified cells are relevant for ERBB2 point mutations. Understanding the sensitivity of distinct ERBB2 variants to the variety of available therapeutics is an active area of investigation^82–84.

In addition to ERBB2 mutations and amplifications, alterations in other members of the receptor tyrosine kinase (RTK) family have also been associated with endocrine resistance. For instance, amplification of EGFR has been reported to confer an endocrine-resistant phenotype, which can be reversed by combining anti-oestrogen with an EGFR inhibitor⁴⁹. Fibroblast growth factor receptor 1 (FGFR1) has also been implicated in preclinical and clinical studies as an important driver of endocrine therapy resistance, both intrinsically and as an acquired mechanism^71,85. Interestingly, in the context of early breast cancer, co-amplification of FGFR1 with an amplicon that contains CCND1 (encoding the ER target gene cyclin D1) together with the FGFR ligands fibroblast growth factor 3 (FGF3), FGF4 and FGF19 was found to be enriched in patients whose tumours maintained a high degree of proliferation following treatment with an aromatase inhibitor⁷¹.

RTK activation promotes oncogenic signalling commonly through both the PI3K–AKT pathway and the RAS–RAF–mitogen-activated protein kinase (MAPK) pathway. Notably, activation of either pathway downstream of RTK activation can also induce breast cancer cell growth independent of ER signalling^22,86. The MAPK pathway can be activated by either loss-of-function mutations in negative regulators of the pathway (for example, NF1) or activating mutations in other pathway components (for example, KRAS, BRAF and MAP2K1), and both types of alterations have been associated with resistance to endocrine therapy^49,87,88. As with ERBB2 activating mutations, NF1 loss-of-function mutations were more than twice as common in tumours previously treated with endocrine therapy when compared with pre-therapy samples and are mutually exclusive with mutations in ESR1. This means that NF1 and ESR1 mutations do not tend to occur in the same tumours, suggesting that they are independent mechanisms of acquired resistance. KRAS and BRAF mutations are similarly exceedingly rare in primary ER⁺ breast cancers and become more frequently detected in therapy-exposed tumours, while remaining at an overall low prevalence⁴⁹.

Overall, the weight of evidence from preclinical functional studies, together with clinical genomic studies, suggests that activation of RTKs and their downstream signalling effectors can support proliferation of ER⁺ breast cancer cells independently of ER activity and thus represents an important theme in therapeutic resistance. Conversely, inhibiting these oncogenic signals without inhibiting ER signalling limits efficacy, and so combinatorial treatments have been the mainstay of these studies. What remains a major challenge for categorizing these mechanisms, quantifying their overall abundance in different clinical scenarios, and developing strategies to overcome such resistance, is the various ways these overlapping pathways are activated.

As a distinct resistance theme, alterations in transcriptional regulators beyond ER have recently emerged as a potential mechanism for endocrine resistance. Amplifications and hotspot mutations in both the transcription factors MYC and CTCF were found more frequently in tumours that have progressed after endocrine therapy in the metastatic setting, especially in those without an ESR1 mutation or MAPK pathway alteration, suggesting that they may be independent drivers of endocrine resistance⁴⁹. In addition, mutations in the transcription factor FOXA1, a protein that cooperates with ER as a pioneer factor, were associated with poorer outcomes for patients with ER⁺ breast cancer treated with aromatase inhibitors. FOXA1 alterations were found mutually exclusively with ESR1 mutations, meaning that they tended to occur in ESR1 wild-type tumours. In the case of mutations in the Wing2 region of FOXA1, this was thought to be related to an increase in FOXA1 chromatin occupancy at ER binding sites and thus enhanced ER-mediated transcription. Another FOXA1 mutation, SY242CS, however, appeared to activate an alternative transcriptome by binding to non-canonical DNA motifs⁸⁹. Although the transcriptional consequences of different classes of FOXA1 mutations (Wing2 versus SY242CS) are distinct, functional experiments showed that both types of variants may promote oestrogen independence, consistent with resistance to aromatase inhibitors. However, they retained sensitivity to the SERM tamoxifen and the SERD fulvestrant, suggesting that overall dependency on ER signalling is likely maintained. Furthermore, in the developing mammary gland, FOXA1 directs lineage programming towards the luminal ER⁺ cell fate, and hormonal therapy resistance in this context may be related to an alteration in cell lineage^90,91. Indeed, lineage plasticity is increasingly recognized as an important mechanism of therapy resistance in other types of cancer, including both prostate cancer under pressure of AR-directed therapies and lung cancer under the pressure of EGFR-targeted therapies^92,93. The role of lineage plasticity in breast cancer hormonal therapy resistance is less well defined. However, approximately 10–20% of metastatic ER⁺ breast tumours appear to exhibit a loss or decrease in ER expression^94–96. Several epigenetic mechanisms have been proposed to account for this change, including abnormal DNA methylation of the ER CpG island⁹⁷, chromatin inactivation by histone deacetylation⁹⁸ or inhibition of ER transcription by the transcription factor Twist⁹⁹. Loss of ER expression as part of a mechanism of resistance likely reflects a fundamental reshaping of the transcriptome, cell identity and tumour phenotype. Sensitivity of ER⁺ luminal cells to endocrine therapy is underpinned by the requirement for ER signalling to drive cell cycle progression in luminal cells specifically. Thus, evolution to alternative cell identities likely comes with reduced reliance on ER signalling for proliferation and consequently resistance to endocrine therapy.

Overcoming therapeutic resistance

The frequent emergence of ESR1 mutations in endocrine therapy-exposed ER⁺ breast cancers suggests that a large fraction of endocrine therapy resistance is related to the particular features and limitations of today’s endocrine therapies, rather than an acquired abnegation of ER function. There has thus been significant energy and investment put towards the pursuit of fully optimized ER-targeted therapies to address refractory metastatic disease. Of the current commonly used anti-oestrogen agents, the SERD fulvestrant is capable of suppressing the activity of mutant ER in vitro, albeit with reduced potency relative to wild-type ER. This is unfortunately the case with all ER antagonists, as they engage with the LBD^41,46. Efforts to develop next-generation endocrine therapies have therefore prioritized the development of molecules that have a fulvestrant-like SERD mechanism of action, but are also orally bioavailable and can be dosed to reach in-patient drug concentrations that will be capable of achieving full antagonism of mutated ER¹⁰⁰.

There have been a plethora of ER binding ligands explored for their potential as next-generation endocrine agents. These ligands have various core structures that drive molecular interactions within the ligand-binding pocket, together with a number of different side chains which protrude from the pocket and differentially impact the positioning of helix 12 of the LBD. As described earlier, the positioning of helix 12 is critical for the recruitment of co-activator proteins to the ER LBD and therefore to the overall activity of ER as a transcription factor. These ligands can be loosely categorized as having either an acrylic acid side chain (for example, etacstil (GW5638, active metabolite GW7604), GDC-0810, AZD9496, rintodestrant (G1T48) and LSZ102) or a basic side chain (for example, elacestrant (RAD1901), GDC-0927, giredestrant (GDC-9545), amcenestrant (SAR439859) and camizestrant (AZD9833))^101,102. GW5638, the first compound with an acrylic acid side chain, is a tamoxifen analogue initially synthesized as an alternative to hormonal therapy in post-menopausal women to prevent osteoporosis¹⁰³. GW5638 and its active metabolite GW7604 were found, similar to tamoxifen, to displace helix 12 from the co-activator binding groove and thus inhibit ER-mediated transcription. In addition to blocking the recruitment of ER co-activators, distinct from tamoxifen, these compounds were proposed to also disrupt the ER surface charge, leading to ER ubiquitylation and proteasomal degradation^104,105. Despite somewhat promising preclinical data, results from early-phase clinical trials of SERDs with an acrylic acid side chain, thought to act similar to GW5638, have been disappointing, with inconsistent response rate and a significant gastrointestinal toxicity profile^{9,68,100,102,106}. Although the acrylic acid-containing SERDs lead to ER degradation and exhibit a fulvestrant-like full ER antagonist profile in MCF-7 cells, they exhibit partial agonist activity in other ER⁺ breast cancer cell lines and also in the uterus, leading to their redesignation as SERM/SERD hybrids^20,107. One hypothesis is that the SERM-like, partial ER agonist properties of these molecules — as is the case with SERMs — limit their efficacy relative to what is achieved by more fully ER-inhibitory agents. Indeed, the key driver behind the discovery and development of fulvestrant was its lack of partial agonism, with the hypothesis that full ER antagonists would be more efficacious than those exhibiting context-dependent ER agonism¹⁶.

Compounds containing basic side chains display a more consistent full ER antagonist profile and have demonstrated robust on-target activity in both ESR1 wild-type and mutant models^{101,108–111}. Several clinical trials have been conducted or are ongoing to evaluate the efficacy of these agents, both alone and in combination with CDK4/CDK6 or PI3K inhibitors, demonstrating mixed results^68,101. For instance, despite strong preclinical and early clinical trial data for amcenestrant, the phase II trial AMEERA-3 (NCT04059484)¹¹², which compared single-agent amcenestrant with physician’s choice anti-oestrogen therapy, failed on its primary end point of progression-free survival (PFS). Notably, 79% of the patients in this trial received previous CDK4/CDK6 inhibitors and 90% received fulvestrant in the physician’s choice arm, which might have complicated the interpretation of amcenestrant efficacy^113,114. The phase III trial, AMEERA-5 (NCT04478266)¹¹⁵, aimed to evaluate the combination of amcenestrant and palbociclib (a CDK4/CDK6 inhibitor) in comparison to palbociclib and letrozole (an aromatase inhibitor), in advanced ER⁺ breast cancer; this trial also ended when it did not meet the prespecified efficacy boundary for continuation^114,116. The clinical development of amcenestrant has subsequently been discontinued.

By contrast, elascetrant became the first oral SERD to demonstrate a statistically significant improvement in PFS in patients with ESR1-mutant tumours¹¹⁷ on the basis of the results of the EMERALD trial (NCT03778931)¹¹⁸. In January 2023, elacestrant received FDA approval for the treatment of post-menopausal women with advanced or metastatic ER⁺HER2⁻, ESR1-mutated breast cancer. The contrasting termination of amcenestrant development versus the success of elacestrant in demonstrating benefit and gaining regulatory approval illustrates both the challenge, and the importance, of optimizing multiple molecule parameters (including potency, pharmacokinetics and drug interaction properties) and clinical trial design in the drug development process. Amcenestrant and elacestrant exhibit similar potencies against ER⁺ breast cancer cells in vitro, and exposure differences do not obviously explain a superiority of elacestrant over amcenestrant. Different patient populations may thus have played some role in the different trial outcomes, but unique attributes of the molecules themselves cannot be ruled out as potential contributing factors. Camizestrant in the SERENA-2 trial (NCT04214288)¹¹⁹ and giredestrant in the acelERA (NCT04576455) trial¹²⁰ both demonstrated promising clinical activity against ESR1-mutant tumours, relative to fulvestrant. These molecules also exhibit significantly higher potency than amcenestrant in preclinical studies^121–123. Additionally, the drug–drug interaction between amcenestrant and palbociclib required lowering the administered dose of amcenestrant in the AMEERA-5 (NCT04478266)¹¹⁵ trial, whereas camizestrant and giredestrant have not required dose reductions owing to drug–drug interactions. The outcomes of the large series of clinical trials that are currently underway in both early and metastatic ER⁺ breast cancer will enable a more complete understanding of the molecule features most critical for driving improvements over the current standard of care and the patient populations that will gain the most benefit from these next-generation therapeutics. Interestingly, sub-group analysis in the EMERALD (NCT03778931)¹¹⁸, SERENA-2 (NCT04214288)¹¹⁹ and acelERA (NCT04576455)¹²⁰ trials suggested that these agents are beneficial relative to the physician’s choice of endocrine therapy (primarily fulvestrant) mostly in the ESR1-mutant setting. The reasons behind the benefit being restricted to ESR1-mutant populations in these studies may be related to the value of ESR1 mutation as a biomarker for identifying tumours that retain a dependence on ER in the advanced setting. However, the superiority of the new endocrine agents may well manifest in ESR1-wild-type populations where endocrine resistance has not yet emerged. For instance, the coopERA trial (NCT04436744)¹²⁴ showed a superiority of giredestrant versus aromatase inhibitors in treatment-naive early ER⁺ breast cancer¹²⁴. Further transcriptional and genomic profiling of pre-treatment and on-treatment biopsies from several of these trials is currently underway to fully interrogate these hypotheses and potentially develop additional biomarkers for patient selection.

Although the development of orally bioavailable SERDs has taken centre stage in the development of next-generation ER-targeted therapies, other novel strategies to target ER have also been pursued; these include the development of a proteolysis targeting chimaera (PROTAC) and selective oestrogen receptor covalent antagonists (SERCAs). PROTACs (also known as ligand-directed degraders or chemical inducers of degradation) are heterobifunctional small molecules that combine a ligand for the target (in this case, ER), with a ligand for an E3 ligase, thereby recruiting the E3 ubiquitin ligase complex to the protein of interest and targeting it for proteasomal degradation¹²⁵. It has been hypothesized that this unique catalytic-type mechanism of action will allow PROTACs to achieve a particularly high degree of selectivity and potency. This is due to their ability to induce energetically favourable protein–protein interactions between their targets and E3 ligase, to recycle each drug molecule through multiple rounds of activity and to refine binding specificity to target proteins¹²⁶. The first ER-PROTAC to enter clinical trial, ARV-471, showed a favourable tolerability profile and evidence of clinical activity, including in patients previously treated with fulvestrant, CDK4/CDK6 inhibitors or oral SERDs¹²⁷. The first-in-class SERCA, H3B-5942, is unique in its ability to bind to the ER LBD covalently and irreversibly via a cysteine at residue 530 of ER (versus the other ligands that display reversible binding to the same domain). This outcompetes oestrogen for binding to the LBD and prevents recruitment of co-activator proteins. SERCAs can inhibit both wild-type and mutant ER, albeit with reduced potency for mutant ER, similar to all other LBD binders. In preclinical studies, H3B-5942 demonstrates similar antiproliferative potency to fulvestrant in vitro, but greater activity versus fulvestrant in vivo¹²⁸. In anticipation of the emergence of mutations at the C530 site that may render cells resistant to H3B-5942, a second SERCA, H3B-6545, was subsequently designed and shown to maintain in vitro potency even in the presence of a mutation at C530 (ref. 129). Notably, covalent antagonism of the LBD does not preclude partial ER agonism, and H3B-6545 exhibits tamoxifen-like ER stimulation in the bone and uterus. This differentiates the signalling properties of SERCAs from the more ER-inhibitory SERDs. In a phase II clinical trial (NCT03250676)¹³⁰, H3B-6545 demonstrated evidence of clinical activity in heavily pretreated patients and those with ESR1 mutations^131,132. Although both ARV-471 and H3B-6545 have demonstrated evidence of being clinically active, larger randomized trials with an approved endocrine therapy comparator will be required to determine whether these alternative mechanisms of ER antagonism are advantageous relative to currently approved agents.

Combining anti-oestrogen and targeted therapies

Another strategy to prevent the emergence of endocrine therapy resistance is to treat tumours with rational drug combinations that co-target the oncogenic signals that cooperate with ER to drive growth. Various approaches have been tested both preclinically and clinically but none more successfully than inhibition of CDK4 and CDK6 in combination with ER suppression. Indeed, this approach has now proven to confer PFS and overall survival benefits in both the early-stage and metastatic settings^133,134.

As described earlier, ER and the CDK4/CDK6:D-cyclin complex converge on regulating the G1-S cell cycle checkpoint. The approved CDK4/CDK6 inhibitors palbociclib, abemaciclib and ribociclib suppress CDK4 and CDK6 activity and thereby attenuate the phosphorylation of RB1, which then acts as a break on the cell cycle by preventing the activity of E2F transcription factors¹³⁵. The introduction of CDK4/CDK6 inhibitors into routine clinical practice was anticipated to generate resistance through loss-of-function RB1 alterations, resulting in a loss of the cell cycle ‘break’ and rendering both CDK4/CDK6 and ER inhibition irrelevant. Interestingly, however, RB1 mutations have remained relatively uncommon, with ESR1 mutations occurring more frequently. For instance, analysis of paired baseline and end-of-treatment circulating tumour DNA sequencing from patients in the PALOMA-3 trial (NCT01942135)¹³⁶ reveals that, as expected, RB1 mutations emerged specifically in the treatment arm that combined the CDK4/CDK6 inhibitor palbociclib with the SERD fulvestrant; however, it was found only in a small number of patients overall at the end of treatment (4.7%). By contrast, a far greater proportion (31.3%) of patients in both the combination arm and the fulvestrant alone arm was found to have an ESR1 mutation at the end of treatment. Nevertheless, acquired resistance to combined inhibition of ER and CDK4/CDK6 can develop through genomic alterations that are more relevant for the combination than for the single-agent endocrine therapy. For instance, cells may become resistant to CDK4/CDK6 inhibitors by reactivating CDK4/CDK6 kinase activity itself, often via induction of the expression of CDK6, spawning the development of next-generation inhibitors of CDK4 and CDK6 as well as inhibitors of downstream CDK2 (refs. 137–140).

In addition to CDK4/CDK6 inhibitors, agents targeting oncogenic signals that function in parallel to ER have also been evaluated to address the challenges of endocrine resistance. In particular, the PI3K–AKT–mTOR pathway has been demonstrated to be mutationally activated in upwards of 50% of ER⁺ breast cancers; thus, agents targeting PI3K, AKT and mTORC1 have been combined with anti-oestrogens and shown meaningful improvements in clinical trials^141–143. Agents in this class have suffered challenges with the toxicities of targeting this pathway with sufficient potency to cause complete pathway inhibition. To this end, newer approaches such as mutant-allele specific inhibitors are now under investigation. In addition, in tumours in which activation of RTK and MAPK signalling may mediate endocrine resistance, as discussed earlier, inhibitors of those pathways (for example, inhibition of EGFR, HER2, FGFR and MEK/ERK) have the potential to augment therapy response^{49,82–84,144}. A major hurdle in bringing this concept to fruition is the fragmentation of the patient population across many distinct genetic variants, each of which requires a specific therapeutic targeting approach. These are extremely challenging trials to execute, particularly for therapeutic agents that also face significant tolerability challenges. Adding to this complexity, multiple mechanisms of resistance can also occur in the same patient. A salient example comes from a research autopsy analysis of multiple metastases from a patient with ER⁺ breast cancer, which revealed that different mechanisms of resistance were present in different metastases. Specifically, ESR1 and ERBB2 mutations, each sufficient to mediate endocrine resistance, were found in distinct metastatic sites. Perhaps even more telling, cell-free DNA analyses have demonstrated the presence of subclonal levels of several different ESR1 mutations in the same patient^48,145. Taken together, this suggests that multiple mechanisms of resistance that may appear mutually exclusive in large sequencing cohorts may in fact be co-occurring in the same patient across different subclones.

As an additional opportunity for combination approaches, other nuclear receptors are known to cooperate with ER in both mammary phenotypes and tumour growth. For instance, progesterone receptor (PR) is highly expressed in the majority of ER⁺ breast cancers and is both a direct transcriptional target of ER and a modulator of ER activity^146,147. Early studies of PR inhibition in ER⁺ PR⁺ tumours were limited by the availability of PR-specific compounds, but newer agents such as onapristone have shown some signal of benefit in heavily treated PR⁺ malignancies and may potentiate the effects of anti-oestrogen compounds^148,149. Interestingly, agonism of PR has also been studied and found to exert anti-proliferative effects in combination with ER antagonists¹⁴⁶. In fact, ample historical data spanning a few decades demonstrates that medroxy-progesterone acetate and its analogue megesterol acetate (Megace) are effective agents with comparable benefits to tamoxifen as a first-line therapy in ER⁺ breast cancer, as has been previously reviewed here¹⁵⁰. The complex role of progesterone, as well as the potential for both agonists and antagonists of progesterone to function as therapeutic agents in the setting of endocrine resistant breast cancer, thus remains to be elucidated.

Similarly, the androgen receptor (AR) is expressed in the vast majority of ER⁺ breast cancers^151–153. In the context of hormone-therapy-resistant breast cancer, AR inhibition with enzalutamide was proposed to inhibit oestrogen-mediated tumour growth in ER⁺ AR⁺ models and to act synergistically with inhibitors of oestrogen^154,155. One randomized trial (NCT02007512)¹⁵⁶ of the addition of enzalutamide to aromatase inhibition was evaluated and did not provide evidence of clinical improvement in PFS. However, a biomarker-selected group with high levels of AR mRNA showed a suggestion of benefit¹⁵⁷. Offering an alternative hypothesis, recent work suggests that AR can act as a tumour suppressor in the context of ER⁺ breast cancer by opposing ER effects on selected promoters. As such, AR agonism with selective AR modulators has been proposed as a viable therapeutic opportunity that is actively being explored in the clinic¹⁵⁸. These apparently opposing therapeutic hypotheses regarding both PR and AR highlight the complexity of hormone receptor interactions and suggest that further study is required to define patient populations within ER⁺ breast cancer that may benefit from therapeutic manipulation of hormone receptors beyond ER itself¹⁴⁶.

Novel treatment modalities

There are perhaps two major challenges to the success of endocrine therapies and their combinations with signalling inhibitors. The first is that these agents frequently achieve cell cycle control without causing extensive cell death, thus allowing an opportunity for escape from therapy. Second, there is extensive heterogeneity both across and within patients with respect to the individual variants that cause therapeutic resistance. The most attractive approaches for transforming patient outcomes in ER⁺ breast cancer will tackle both challenges, leading to elimination of tumour cells and acting independently of specific, often relatively rare, signalling alterations. We propose that antibody–drug conjugates (ADCs) and immunotherapies have the potential to meet these criteria. ADCs have been extremely impactful in the treatment of ERBB2-amplified breast cancers, 50% of which are also ER⁺. In a major recent advance for patients without ERBB2 amplification, the DESTINY-Breast04 (NCT03734029)¹⁵⁹ trial demonstrated a strong improvement in both PFS and overall survival for patients with HER2-low tumours (defined as those scoring 1 or 2 on immuno-histochemistry and a negative fluorescence in situ hybridization) and previous evidence of endocrine therapy resistance when treated with the ADC trastuzumab deruxtecan (T-DXd). This established that T-DXd, which delivers a topoisomerase I inhibitor to HER2-expressing tumour cells, was broadly active even in heavily pretreated ER⁺ metastatic breast cancer¹⁶⁰. Moreover, the phase III trial TROPiCS-02 (NCT03901339)¹⁶¹ also recently established the benefit of the ADC sacituzumab govitecan, which delivers chemotherapy to TROP2-expressing cells, in metastatic ER⁺ HER2⁻ disease¹⁶². This drug likewise received FDA approval for the treatment of advanced metastatic ER⁺ breast cancer in early 2023. These studies represent important advances for the utility of ADCs beyond ERBB2-amplified breast cancer, and their early success has restructured the treatment paradigm for advanced, refractory ER⁺ breast cancer.

Immunotherapy, such as immune checkpoint blockade, has become the cornerstone of treatment in many types of cancer. However, ER⁺ HER2⁻ breast cancer has traditionally been considered ‘immune cold’ with lower numbers of tumour infiltrating lymphocytes¹⁶³ and lower expression of PDL1 (ref. 164). Interestingly, ER signalling itself appears to have an immunosuppressive effect on the tumour microenvironment¹⁶⁵. How and where we may combine endocrine therapy with immunotherapy to take advantage of their potential synergy is thus a topic of much active investigation¹⁶⁶. One novel approach to leveraging the immune system in relatively ‘cold’ tumours might be individualized mRNA neoantigen vaccines. Although cancer vaccines have faced significant challenges in the metastatic setting, a small proof-of-concept study of personalized cancer vaccine autogene cevumeran (NCT04161755)¹⁶⁷ was recently conducted in the adjuvant setting in pancreatic ductal adenocarcinoma¹⁶⁸. Similar to ER⁺ breast cancer, pancreatic ductal adenocarcinoma is refractory to checkpoint blockade, with most tumours exhibiting an immune-excluded or immune-desert phenotype^169–171. Preliminary data suggest that patients who successfully mounted an immune response gained benefit from the individualized vaccine approach. However, it remains to be seen whether this early, but highly intriguing, signal will be validated in larger studies and in other indications, including in ER⁺ breast cancer.

Towards a tumour-evolution informed approach

As we gain a deeper understanding of how ER⁺ breast tumours evolve under the pressure of hormonal therapy to cultivate mechanisms of resistance, we will also need to leverage new technologies to monitor tumour evolution in real time. For instance, liquid biopsies could provide information about response or failure to ongoing treatment and also data on tumour heterogeneity and evolution throughout the course of the disease^172,173. More recently, plasma cell-free DNA was successfully utilized to assess high-resolution ER and FOXA1 DNA-binding profiles, from which ER activity, and therefore endocrine therapy sensitivity, may be inferred¹⁷⁴. This technology would allow for a minimally invasive means to gain insights into tumour signalling and potentially ER dependence in real time, which would significantly enhance our ability to rapidly identify patients who may gain additional benefit from ER-directed therapies versus those who need treatments targeted against non-endocrine oncogenic signals. Although this approach, which measures the binding of a transcription factor to DNA, may have utility for a number of cancers, it is particularly well suited to ER⁺ breast cancer, given that ER is a transcription factor. These and other techniques will become more relevant in real-time clinical monitoring and adaptive treatment of patients with breast cancer in the years ahead.

Conclusions and perspective

ER signalling represents a foundational dependency in ER⁺ breast cancer cells, making ER-targeted therapies the foundation of treatment in this tumour type. The discovery of activating ESR1 mutations, occurring in up to 40% of the patients following exposure to aromatase inhibitors in the metastatic setting, as the major contributor to acquired resistance to commonly used endocrine therapies, suggested that ER dependency is maintained through disease progression and was thus a major spark triggering community-wide investment in further optimized endocrine therapies. These next-generation ER-targeted agents are being rigorously evaluated, with multiple pivotal trials currently playing out across both primary and metastatic breast cancers. The recent FDA approval of elacestrant for the treatment of metastatic ER⁺ HER2⁻ breast cancers harbouring ESR1 mutations, has provided important validation for the hypothesis that there remains an opportunity to bring further benefit to patients through more effective targeting of ER. Although the next-generation endocrine therapies have been designed and developed to tackle the currently described oestrogen-independent ESR1 mutations, one important question is whether these endocrine agents will select for novel mechanisms of ER reactivation, or whether they will skew tumours entirely towards ER independence. The community eagerly awaits the outcomes of ongoing trials, as well as of the genomic and transcriptomic analyses of tumour biopsies and circulating tumour DNA from patients exposed to these novel therapies for prolonged periods of time. These data will represent the leading edge of ER⁺ tumour evolution and set the stage for future therapeutic exploits in this disease.

Glossary

ADCs

A class of drugs that combine monoclonal antibodies with cytotoxic agents to specifically target cells expressing the antigens recognized by those antibodies.

Aromatase inhibitors

A class of drugs that block the synthesis of oestrogens by the aromatase enzyme in non-ovarian tissues.

LBD

The C-terminal ligand-binding domain of ER, responsible for the binding of oestrogen ligand.

Liquid biopsies

A technology that allows sampling of body fluids such as blood for molecular profiling, allowing for a less-invasive way of tumour monitoring, detection and characterization than traditional tissue-based biopsies.

SERCA

A class of anti-oestrogen agents that bind covalently to the LBD of ER and prevent recruitment of co-activators.