Epigenetic mechanisms in breast cancer therapy and resistance

Abstract

The majority of breast cancers express the estrogen receptor (ERα) and agents targeting this pathway represent the main treatment modality. Endocrine therapy has proven successful in the treatment of hormone-responsive breast cancer since its early adoption in the 1940s as an ablative therapy. Unfortunately, therapeutic resistance arises, leading to disease recurrence and relapse. Recent studies increased our understanding in how changes to the chromatin landscape and deregulation of epigenetic factors orchestrate the resistant phenotype. Here, we will discuss how the epigenome is an integral determinant in hormone therapy response and why epigenetic factors are promising targets for overcoming clinical resistance.

Fig. 1. ERα mediates epigenetic changes by interacting and crosstalking with pioneer factors and co-regulators.

a Schematic depiction of various structural domains within the estrogen receptor (ERα). The LBD harbors surfaces for dimerization as well as coactivator binding. Upon binding E2, helix 12 within the LBD shifts to an active conformation, which promotes ERα interaction with coactivators. b Pioneer factors such as FOXA1, GATA3, PBX1, and AP-2γ preferentially bind to hypomethylated genomic sites bearing their respective motifs. Importantly, these sites are often already marked with low levels of H3K4me1/2, which increase with the recruitment of pioneer factors such as FOXA1. Upon stimulation with the E2 ligand, pioneer factors facilitate the localization of liganded ERα to the chromatin. This leads to the activation of gene expression as ERα recruits epigenetic activators such as P300/CBP, the SWI/SNF complex, PRMTs and EZH2 (through direct contact or through coactivators such as SRC-1/2/3 and Mediator) to deposit activating epigenetic marks such as H4R3me1 and H3K27Ac (solid-colored). However, through interactions with corepressors such as LCOR and NCoR1/2, liganded ERα and tamoxifen-bound ERα can also recruit epigenetic repressors including HDACs and the NuRD complex to mediate gene repression by removing active epigenetic marks (faintly colored).

Fig. 2. The development of embryonic mammary glands is dependent on carefully coordinated spatial-temporal activation of signaling pathways such as WNT and Hedgehog (SHH).

In normal mammary epithelium, DKK3 binds to LRP, a WNT pathway coactivator of Frizzled, which prevents the activation of the pathway in the presence of the WNT ligand. E-Cadherin binds to cytoplasmic β-catenin, which is degraded by GSK3β in the absence of WNT activation. The promoter of SHH (encoding the Hedgehog ligand SHH) is hypermethylated and the Hedgehog pathway is silenced. In breast cancer, however, the DKK3 promoter is hypermethylated, which leads to its downregulation. In the absence of DKK3, LRP can coactivate Frizzled in the presence of the WNT ligand, leading to phosphorylation of DSH, which inhibits GSK3β from degrading β-catenin. E-Cadherin is also downregulated via promoter methylation. In addition, the SHH promoter becomes hypomethylated, thereby upregulating the expression of SHH and activating the Hedgehog pathway via GLI1. Activation of the WNT and the Hedgehog pathways lead to stem cell renewal, EMT, metastasis, and tamoxifen resistance.

Fig. 3. Illustrations of selected genomic alterations that mediate therapeutic resistance in ER⁺ breast cancer.

a NF1 is a GTPase that (1) inhibits RAS activation of the MAPK pathway and (2) functions as a corepressor of ERα at the CCND1 gene, which encodes cyclin D1. Loss of NF1 in treatment-resistant cells results in increased activation of the MAPK signaling pathway and overexpression of cyclin D1, which promotes G1/S transition by activating CDK4/6. MEK inhibitors and CDK4/6 inhibitors can overcome MAPK and CDK4/6 overexpression, respectively. Fulvestrant can also be used in cases where there is loss of NF1. b In treatment-sensitive ER⁺ breast cancer, ERα dimerizes upon binding to E2 and is phosphorylated by CDK7. ESR1 mutations often occur within the LBD of resistant cells, leading to constitutively active ERα mutants that mediate gene expression through coactivator interactions independent of E2. Proposed therapeutic strategies include inhibition of CDK7 with THZ1 and inhibition of SRC coactivators. c CYP19A1 codes for aromatase, the enzyme that converts testosterone into estrogen. CYP19A1 overexpression is often acquired in patients that relapse after AI treatment, leading to increased estrogen and treatment resistance. Patients with CYP19A1 overexpression can be treated with irreversible steroidal AIs instead of reversible AIs as well as with fulvestrant. d FOXA1 is often amplified or mutated in treatment-resistant ER⁺ breast cancer, leading to increased FOXA1 activity and redistribution of FOXA1 to de novo enhancers. These de novo sites are enriched near TF motifs such as AP-1, STAT5, and SOX9, and promote a metastatic transcriptional program. Targeting FOXA1 directly is challenging. However, specific inhibitors can target FOXA1 downstream genes that mediate metastasis, such as HIF-2α.

Fig. 4. Schematic depiction of various dysregulated epigenetic pathways in treatment- resistant ER⁺ breast cancer that are potential targets for novel epigenetic therapeutics.

Blue and yellow panels: the PI3K/AKT pathway can phosphorylate DNMT1, which stabilizes it on the chromatin, leading to maintenance of DNA hypermethylation. EZH2 is also phosphorylated by PI3K/AKT, which depletes H3K27me3 genome-wide. In addition, KMT2D phosphorylation by PI3K/AKT depletes H3K4me1/2, which decreases FOXA1 (and therefore, also ERα) chromatin binding, leading to hormone therapy resistance. These aberrant epigenetic pathways can be targeted by PI3K inhibitors. Furthermore, DNMT1 stabilization and EZH2 inhibition can be targeted by DNA hypomethylating agents and HDAC inhibitors, respectively. Purple panel: HDAC recruitment to the ESR1 promoter leads to reduced H3K27ac, which results in DNMT1-mediated promoter hypermethylation and drug resistance via ESR1 downregulation. HDAC inhibition (entinostat) can be used to reverse HDAC-mediated ESR1 downregulation. Orange panel: KDM5 (KDM5A/B) is a family of histone H3 lysine 4 demethylases associated with therapeutic resistance in different cancer types. Increased activity of KDM5 enzymes leads to reduction in H3K4me3 levels and, as a result, increased transcriptional heterogeneity. Particularly, high KDM5B expression levels are associated with poor prognosis in ER⁺ breast cancer. Inhibitors to modulate the activity of KDM5 family members can improve the response to endocrine therapy agents such as fulvestrant. Green panel: loss of KMT2C redistributes ERα to AP-1-regulated genes to promote hormone-independent but ERα-dependent transcription, suggesting that treatment with SERDs (fulvestrant) and CDK4/6 inhibition may be viable therapeutic options. Red panel: ARID1A is a component of the SWI/SNF complex, and its mutation leads to limited chromatin accessibility for ERαα and FOXA1 at genes that regulate luminal cell fate, as well as promotes a switch from a luminal phenotype to a basal-like phenotype. ARID1A mutations can be targeted with EZH2, HDAC, BET, and PI3K/AKT inhibitors.