ESR1 ligand-binding domain mutations in hormone-resistant breast cancer

Abstract

Seventy percent of breast cancers express estrogen receptor (ER), and most of these are sensitive to ER inhibition. However, many such tumors for unknown reasons become refractory to inhibition of estrogen action in the metastatic setting. We conducted a comprehensive genetic analysis of two independent cohorts of metastatic ER-positive breast tumors and identified mutations in ESR1 affecting the ligand-binding domain (LBD) in 14 of 80 cases. These included highly recurrent mutations encoding p.Tyr537Ser, p.Tyr537Asn and p.Asp538Gly alterations. Molecular dynamics simulations suggest that the structures of the Tyr537Ser and Asp538Gly mutants involve hydrogen bonding of the mutant amino acids with Asp351, thus favoring the agonist conformation of the receptor. Consistent with this model, mutant receptors drive ER-dependent transcription and proliferation in the absence of hormone and reduce the efficacy of ER antagonists. These data implicate LBD-mutant forms of ER in mediating clinical resistance to hormonal therapy and suggest that more potent ER antagonists may be of substantial therapeutic benefit.

Conflict of interest statement

Competing Financial interests

The authors declare no competing financial interests.

Figures

Figure 1. ESR1 mutations in ER+ metastatic breast cancers

a. Percentage of matched cases with gene mutations detected in our patient samples (MSKCC) were compared to those of The Cancer Genome Atlas (TCGA) invasive breast cancer and luminal A/B cases. Only genes found mutated in 3 or more cases are shown. b. Top amplified genes detected in the patient samples. Genes were considered amplified when the ratio of their copy number in tumor to normal is greater than 2 c. Diagram of ER domains with the location of the identified mutations. AF-1: Activation Function-1; DBD: DNA binding domain; NLS: Nuclear localizing signal; LBD: Ligand binding domain; AF-2: Activation function-2.

Figure 2. ERα LBD mutants demonstrate elevated activity in absence of hormone

a. SKBR3 and MCF7 cells were transfected with empty vector, HA-ERα wild type (WT) or mutants, ERE-luciferase and Renilla luciferase reporter constructs in hormone-depleted medium with 10 nM of E2 added for 24 hours where indicated. Relative light units (RLU) were calculated as the ratio of Firefly luciferase over Renilla luciferase activity. Graph was plotted with the mean ± SD of triplicate biological replicates. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. b. MCF7 cells were transfected with HA-ERα wild type (WT) or mutants in hormone-depleted medium and harvested 48 hours post-transfection. The mRNA level of ERα target genes were quantified by qRT-PCR using gene-specific primers. Bars represent mean ± SD of triplicate technical replicates normalized to ESR1 expression. *, P < 0.05; ***, P < 0.001; ****, P < 0.0001. The data shown is a representative of two biological replicates. c. Gene expression microarray was performed using RNA derived from MCF7 cells transfected with empty vector, GFP-ERα WT or mutants in hormone-depleted medium for 48 hours. A total of 92 genes were commonly up or down-regulated by the ERα mutants and they were hierarchically clustered to generate a heat map. Selected genes previously well established as estrogen responsive genes were marked with *. d. Lysates of MCF7 cells transfected with empty vector, HA-ERα WT or mutants in hormone-depleted medium for 48 hours, were immunoblotted with phospho-Serine 118 of ERα, HA and actin antibodies. The immunoblots shown are a representative of experiments repeated at least three times. e. Lysates generated from MCF7 transfected with empty vector GFP-ERα WT, or D538G were immunoprecipitated with GFP-Trap beads and immunoblotted with AIB1 and GFP antibodies. The immunoblots shown are a representative of experiments repeated three times. f. MCF7 cells transfected with HA-ERα WT or mutants were treated with SNX2112 at various doses for 3 hours before harvest. Lysates were then immunoblotted with HA and actin antibodies. The immunoblots shown are a representative of experiments repeated three times. g. Mice bearing MCF7 inducible vector control (VC), HA-ER WT or Tyr537Ser or Asp538Gly tumors had their 17β-estradiol pellets removed when the tumors reached an approximate volume of 250 mm3 and then fed with water containing 0.2–0.5 mg/ml of doxycycline a week after for the induction of ER WT and mutant expression. The result was presented as percentage tumor growth by normalizing to the first mean tumor volume measured for each group ± SEM (n = 10 mice/group).

Figure 3. ER antagonists partially inhibit ER mutant transcriptional activity

SKBR3 cells were transfected with HA-ERα WT or mutants, ERE-luciferase and Renilla luciferase reporter constructs in hormone-depleted medium. The cells were then treated with either a. 1 nM E2 alone or in combination with 4-OHT (4-hydroxytamoxifen) and b. fulvestrant at the indicated doses for 24 hours before luciferase activity measurement. Relative light units (RLU) were calculated as the ratio of Firefly luciferase over Renilla luciferase activity. Graph was plotted with the mean ± SD of triplicate biological replicates. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 4. Molecular dynamics-modeled structures of WT and mutant forms of ERα

a. WT (cyan), b. Tyr537Ser (green), and c. Asp538Gly (red) at 10 ns, overlaid with a co-crystallized structure (PDB accession code 1GWQ) of ERα WT in an agonist conformation (gray) and TIF2 (wheat). d. Hydrogen-to-acceptor distances between Asp538/Ser537/Gly538 and Asp351 in ERα WT/Tyr537Ser/Asp538Gly respectively, over the course of 10-ns molecular dynamics simulations.