A kinetic model identifies phosphorylated estrogen receptor-α (ERα) as a critical regulator of ERα dynamics in breast cancer

Abstract

Receptor levels are a key mechanism by which cells regulate their response to stimuli. The levels of estrogen receptor-α (ERα) impact breast cancer cell proliferation and are used to predict prognosis and sensitivity to endocrine therapy. Despite the clinical application of this information, it remains unclear how different cellular processes interact as a system to control ERα levels. To address this question, experimental results from the ERα-positive human breast cancer cell line (MCF-7) treated with 17-β-estradiol or vehicle control were used to develop a mass-action kinetic model of ERα regulation. Model analysis determined that RNA dynamics could be captured through phosphorylated ERα (pERα)-dependent feedback on transcription. Experimental analysis confirmed that pERα-S118 binds to the estrogen receptor-1 (ESR1) promoter, suggesting that pERα can feedback on ESR1 transcription. Protein dynamics required a separate mechanism in which the degradation rate for pERα was 8.3-fold higher than nonphosphorylated ERα. Using a model with both mechanisms, the root mean square error was 0.078. Sensitivity analysis of this combined model determined that while multiple mechanisms regulate ERα levels, pERα-dependent feedback elicited the strongest effect. Combined, our computational and experimental results identify phosphorylation of ERα as a critical decision point that coordinates the cellular circuitry to regulate ERα levels.

Keywords: feedback; mathematical modeling; nuclear receptor; systems biology.

Figure 1.

An empirically fit model of ERα dynamics captured experimental data trends. A) ESR1 and ERα levels were dynamic after E2 treatment. qRT-PCR was used to analyze the expression of nascent ESR1 (nRNA) and fully processed messenger ESR1 (mRNA) in MCF-7 cells following treatment with 10 nM E2 (34). tERα was measured by pulse-chase analysis (35), and pERα was measured by Western blot as described in the Materials and Methods. Data for nRNA, mRNA, and tERα were normalized to levels of EtOH-treated controls. Because of the experimental method used, tERα measurements included both nonphosphorylated and phosphorylated forms of ERα. Data for pERα were normalized to the initial concentration of tERα. n = 3 for nRNA, mRNA, and pERα; n = 1 for tERα. B) Diagram of reactions included in the empirically fit model of ESR1 and ERα following E2 treatment. C) Comparison of the empirically fit model to experimental data. In all figures, model results are represented by solid lines and experimental data are represented by circles.

Figure 2.

Inclusion of pERα negative feedback captured RNA trends qualitatively. A) Diagram of the reactions included in the feedback model of ERα regulation, where the red arc represents pERα-dependent negative feedback on ESR1 transcription. B) Comparison of the feedback model fit to experimental data with k_d,3 equal to k_d,2.

Figure 3.

ChIP confirmed the possibility of pERα-dependent negative feedback. Experimental analysis by ChIP demonstrated that pERα-S118 bound directly to the ESR1 promoter, and this interaction significantly increased after 30 min of E2 treatment. *Significantly different (P < 0.05) from EtOH-treated vehicle control, n = 5.

Figure 4.

Increasing pERα degradation rate (k_d,3) captured protein dynamics. Model simulated time courses of nRNA, mRNA, tERα, and pERα following E2 treatment for increasing values of k_d,3 (1–10 times greater than k_d,2).

Figure 5.

Incorporating both negative feedback and increased k_d,3 captured qualitative and quantitative data trends. Comparison of the feedback model to experimental data, with k_d,3 included as a free parameter in model fitting.

Figure 6.

Increases in a were predicted to have the greatest impact on decreasing ERα levels. Sensitivity analysis demonstrated that increasing k₃, k_d,3, and a resulted in a decrease in steady-state tERα, whereas decreasing these parameters resulted in an increase in steady-state tERα.