Androgen receptor promotes tamoxifen agonist activity by activation of EGFR in ERα-positive breast cancer

Abstract

Tamoxifen (Tam) resistance represents a significant clinical problem in estrogen receptor (ER) α-positive breast cancer. We previously showed that decreased expression of Rho guanine nucleotide dissociation inhibitor (Rho GDI) α, a negative regulator of the Rho GTPase pathway, is associated with Tam resistance. We now discover that androgen receptor (AR) is overexpressed in cells with decreased Rho GDIα and seek to determine AR's contribution to resistance. We engineered ERα-positive cell lines with stable knockdown (KD) of Rho GDIα (KD cells). Resistance mechanisms were examined using microarray profiling, protein-interaction studies, growth and reporter gene assays, and Western blot analysis combined with a specific AR antagonist and other signaling inhibitors. Tam-resistant tumors and cell lines with low Rho GDIα levels exhibited upregulated AR expression. Microarray of Rho GDIα KD cells indicated that activation of EGFR and ERα was associated with Tam treatment. When AR levels were elevated, interaction between AR and EGFR was detected. Constitutive and Tam-induced phosphorylation of EGFR and ERK1/2 was blocked by the AR antagonist Enzalutamide, suggesting that AR-mediated EGFR activation was a mechanism of resistance in these cells. Constitutive ERα phosphorylation and transcriptional activity was inhibited by Enzalutamide and the EGFR inhibitor gefitinib, demonstrating that AR-mediated EGFR signaling activated ERα. Tam exhibited agonist activity in AR overexpressing cells, stimulating ERα transcriptional activity and proliferation, which was blocked by Enzalutamide and gefitinib. We describe a novel model of AR-mediated Tam resistance through activation of EGFR signaling leading to ER activation in ERα-positive cells with low expression of Rho GDIα.

Keywords: Androgen receptor; Breast cancer; Endocrine therapy resistance; Epidermal growth factor receptor; Estrogen receptor; Tamoxifen.

Conflict of interest statement

Conflict of interest:

The authors declare that they have no financial conflict of interest.

Figures

Figure 1. Low Levels of Rho GDI Are Associated With AR Overexpression and Activation of EGFR and MAPK/ERK Pathways in TamR Cells

a. Heatmap of selected gene expression patterns from microarray analysis of primary breast cancer tissue from patients whose tumors were Tam-sensitive (TS, n=4) and metastatic breast cancer tissue from patients whose tumors were Tam-resistant (TR, n=5). b. Statistical analysis of ARHGDIA (Rho GDI ) and AR expression from tumors in (A) (Columns=mean, Bars=SD, *p<0.05). c. Immunoblot analysis (left) of Rho GDIα, AR, EGFR, and GAPDH (loading control) in ZR-75-B and MCF-7 cells expressing Rho GDIα shRNA (KD) and empty vector control cells (Vec). Quantification (right) of relative pEGFR levels normalized to loading control (GAPDH). Data represents average of three independent experiments. (Columns=mean, Bars=SD) d. qRT-PCR analysis of ARHGDIA (Rho GDI ) and AR mRNA expression in ZR-75-B and MCF-7 Vec and KD cells (Columns=mean, Bars=SD, *p<0.05). e. Heatmap of selected protein expression patterns as measured by RPPA in MCF-7 Vec and KD cells. Cells were treated with 100nM Tam for 24 hours prior to analysis. Three independent lysates were measured for each cell line and each lysate is represented by a column of the heatmap. f. Statistical analysis of AR and MEK1 protein expression from RPPA in (E) (Columns=mean, Bars=SD, *p<0.05)

Figure 2. Enhanced interaction between AR and EGFR in cells with Rho GDI loss

a. Immunocytochemistry assay showing localization of AR and EGFR in ZR-75-B Vec and KD cells after treatment with vehicle control (veh. ctrl.) (DMSO) or Enz (10μM) for 24 hours. b. Proximity ligation assay showing AR and EGFR complexes in ZR-75-B Vec and KD cells following 24 hour treatment with veh. ctrl. (DMSO) or the AR inhibitor Enzalutamide (Enz) (10μM). DAPI nuclear staining is indicated in blue and PLA reaction products are indicated as red dots.

Figure 3. AR activates EGFR signaling in cells with low Rho GDIα expression

a. Immunoblot analysis (left) of basal pEGFR (tyrosine 845) and GAPDH (loading control) expression in ZR-75-B Vec and KD cells. Corresponding total protein levels are also shown. Samples were run on a single gel; non-adjacent lanes are indicated by black line. Quantification (right) of relative pEGFR levels normalized to loading control (GAPDH) and total protein levels. Data represents average of three independent experiments. (Columns=mean, Bars=SD) b. Immunoblot analysis (left) of pEGFR (tyrosine 845) and GAPDH (loading control) expression in ZR-75-B Vec and KD cells following treatment with the EGFR inhibitor gefitinib (gef) (1μM), or Enz (10μM) for 30 minutes. Corresponding total protein levels are also shown. Samples were run on a single gel; non-adjacent lanes are indicated by black line. Quantification (right) of relative pEGFR levels normalized to loading control (GAPDH) and total protein levels. Data represents average of three independent experiments. (Columns=mean, Bars=SD) c. Immunoblot analysis (left) of basal p-P44/42 MAPK (threonine 202/tyrosine 204) levels in ZR-75-B Vec and KD cells. Corresponding total protein levels are also shown. Samples were run on a single gel; non-adjacent lanes are indicated by black line. Quantification (right) of relative p-P44/42 MAPK levels normalized to loading control (GAPDH) and total protein levels. Data represents average of three independent experiments. (Columns=mean, Bars=SD) d. Immunoblot analysis (left) of p-P44/42 MAPK (threonine 202/tyrosine 204) in ZR-75-B Vec and KD cells following treatment with EGF (200ng/ml), gef (1μM), or Enz (10μM) for 30 minutes. Corresponding total protein levels are also shown. Quantification (right) of relative p-P44/42 MAPK levels normalized to loading control (GAPDH) and total protein levels. Data represents average of three independent experiments. (Columns=mean, Bars=SD) e. Immunoblot assay (left) indicating pEGFR (tyrosine 845) levels of ZR-75-B KD cells following 30 minutes treatment with Tam (100nM) or combination treatment with Tam and inhibitors of AR and EGFR (Enz (10μM) and gef (1μM) respectively). Corresponding total EGFR levels are also shown. Samples were run on a single gel; non-adjacent lanes are indicated by black line. Quantification (right) of relative pEGFR levels normalized to loading control (GAPDH) and total protein levels. Data represents average of three independent experiments. (Columns=mean, Bars=SD, *p<0.05) f. p-P44/42 MAPK (threonine 202/tyrosine 204) levels (left) in ZR-75-B KD cells following 24 hour treatment with Tam (100nM) or combination treatment with Tam and inhibitors of AR and EGFR (Enz (10μM) and gef (1μM) respectively). Samples were run on a single gel; non-adjacent lanes are indicated by black line. Quantification (right) of relative p-P44/42 MAPK levels normalized to loading control (GAPDH) and total protein levels. Data represents average of three independent experiments. (Columns=mean, Bars=SD, *p<0.05)

Figure 4. AR and EGFR promote increased constitutive and Tam induced ER transcriptional activity in cells with Rho GDI loss

a. Immunoblot analysis (left) indicating levels of pERα (Ser 118) in ZR-75-B Vec and KD cells following 1 min of treatment with EGF (200ng/ml), gef (1μM), or Enz (10μM). Total ERα protein levels are also shown. Samples were run on a single gel; non-adjacent lanes are indicated by black line. Quantification (right) of relative pERα (Ser 118) levels normalized to loading control (GAPDH) and total protein levels. Data represents average of three independent experiments. (Columns=mean, Bars=SD, *p<0.05) b. ARE-Luciferase reporter assay looking at AR mediated transcriptional activity following 24 hour treatment with the synthetic AR agonist R1881 (10nM) in ZR-75-B Vec and KD cells. Raw luciferase activity was normalized to expression from a co-transfected β-galactosidase (β-gal) reporter plasmid. (*p<0.05, N.S. = non-significant). c. ERE-Luciferase reporter assay looking at ER activity in ZR-75-B Vec and KD cells following 24hr treatment with 10nM R1881. Raw luciferase activity was normalized to expression from a co-transfected β-gal reporter plasmid. (*p<0.05, N.S. = non-significant). d. ERE-Luciferase reporter assay of ER activity in ZR-75-B Vec and KD cells following 24hr treatment with Enz (10μM) or gef (1μM). Raw luciferase activity was normalized to expression from a co-transfected β-gal reporter plasmid. (*p<0.05, N.S. = non-significant). e. Immunoblot analysis (left) of pERα (Ser 118) and total ERα protein levels in ZR-75-B KD cells following treatment with Tam (100nM), or combination treatment with Tam and inhibitors of AR and EGFR (Enz (10μM) and gef (1μM) respectively) for 30 minutes. Samples were run on a single gel; non-adjacent lanes are indicated by black line. Quantification (right) of relative pERα (Ser 118) levels normalized to loading control (GAPDH) and total protein levels. Data represents average of three independent experiments. (Columns=mean, Bars=SD, *p<0.05) f. ERE-Luciferase reporter assay looking at ER activity following 24hr treatment with 100nM Tam, 10μM Enz, 1μM gef, or combinations as indicated. Raw luciferase activity was normalized to expression from a co-transfected β-gal reporter plasmid. (*p<0.05, N.S. = non-significant).

Figure 5. AR and EGFR mediate growth and Tam-stimulated proliferation in ZR-75-B and MCF-7 cells with low Rho GDIα Expression

a. Soft agar growth assay looking at anchorage independent growth in ZR-75-B and MCF-7 Vec and KD cells after 2 weeks of treatment with R1881 (10nM) (*p<0.05, N.S. = non-significant). b. Soft agar growth assay in ZR-75-B and MCF-7 Vec and KD cells analyzing anchorage independent growth following 2 weeks of treatment with 10μM Enz or 1μM gef. (*p<0.05, N.S. = non-significant). c. Soft agar growth assay in ZR-75-B and MCF-7 Vec and KD cells looking at anchorage independent growth following 2 weeks of treatment with Tam (100nM), Enz (10μM), gef (1μM), and combinations as indicated. (*p<0.05, N.S. = non-significant).