Progesterone receptor isoforms, agonists and antagonists differentially reprogram estrogen signaling

Abstract

Major roadblocks to developing effective progesterone receptor (PR)-targeted therapies in breast cancer include the lack of highly-specific PR modulators, a poor understanding of the pro- or anti-tumorigenic networks for PR isoforms and ligands, and an incomplete understanding of the cross talk between PR and estrogen receptor (ER) signaling. Through genomic analyses of xenografts treated with various clinically-relevant ER and PR-targeting drugs, we describe how the activation or inhibition of PR differentially reprograms estrogen signaling, resulting in the segregation of transcriptomes into separate PR agonist and antagonist-mediated groups. These findings address an ongoing controversy regarding the clinical utility of PR agonists and antagonists, alone or in combination with tamoxifen, for breast cancer management. Additionally, the two PR isoforms PRA and PRB, bind distinct but overlapping genomic sites and interact with different sets of co-regulators to differentially modulate estrogen signaling to be either pro- or anti-tumorigenic. Of the two isoforms, PRA inhibited gene expression and ER chromatin binding significantly more than PRB. Differential gene expression was observed in PRA and PRB-rich patient tumors and PRA-rich gene signatures had poorer survival outcomes. In support of antiprogestin responsiveness of PRA-rich tumors, gene signatures associated with PR antagonists, but not PR agonists, predicted better survival outcomes. The better patient survival associated with PR antagonists versus PR agonists treatments was further reflected in the higher in vivo anti-tumor activity of therapies that combine tamoxifen with PR antagonists and modulators. This study suggests that distinguishing common effects observed due to concomitant interaction of another receptor with its ligand (agonist or antagonist), from unique isoform and ligand-specific effects will inform the development of biomarkers for patient selection and translation of PR-targeted therapies to the clinic.

Keywords: cancer; estrogen; hormones; progesterone.

Figure 1. PRA and PRB have isoform-specific cistromes, interactomes, transcriptomes and phenotypic outcomes

A. Overlap between PRA and PRB binding events in T47D cells treated for 45 minutes with 10 nM R5020. B. Top three binding motifs enriched at the binding sites for only PRA, only PRB or overlapping sites for both the receptors. −Log10(P) depicts the significance for the enrichment of a hormone response element in the binding sites of interest. The full list of motifs is provided in Supplementary Table 1. C. Mass spectrometry of immunoprecipitates of PR isoform-specific pull downs in T47D cells treated with estrogen plus R5020 identifies distinct interactomes for PRA and PRB. Pull down with control IgG was used as a background control. Two independent experiments were performed. D. Unsupervised clustering of sample-sample correlation of transcriptomes observed in T47D cells after 12 hours of treatments with 10 nM E2, R5020 or with both the hormones. High correlation (i.e., correlation coefficient 1) between any two samples is displayed in red and low correlation (i.e., correlation coefficient 0) is displayed in blue.

Figure 2. Differential gene expression in patient tumors expressing disproportionate levels of PRA and PRB

A. Dot plots represent PRA and PRB-regulated genes in T47D cells treated with 10 nM R5020 for twelve hours. PRA is a stronger inducer (or repressor) of gene expression than PRB for the genes represented by blue dots. Conversely, PRB is a stronger inducer (or repressor) of gene expression than PRA for the genes represented by red dots. B. Box plots depict ensemble of magnitude of inhibition of gene expression by PRA or PRB. Dot plots and box plots represent union of PRA and PRB-regulated genes. C., D. Changes in C. matrigel invasion and D. cell confluence of ER+/PRA+ and ER+/PRB+ T47D cells in response to treatments with 10 nM R5020. ^*** denotes P-value < 10^-3. E., F. Heatmaps display E. gene expression and F. sample-sample correlation in five patient tumors expression higher PRA versus PRB and six tumors with higher expression of PRB versus PRA. High correlation (i.e., correlation coefficient 1) between any two samples is displayed in red and low correlation (i.e., correlation coefficient 0) is displayed in blue. Data for nine out of 11 tumors was obtained from Rojas et al [44]. Data for two other unpublished tumors (tumors B3 and B5) used in this study was kindly provided by Dr. Claudia Lanari and Dr. Martin Abba.

Figure 3. Differential transcriptomes in xenografts treated with various PR agonists and antagonists

A. T47D xenografts were grown for about 6 weeks and then were subsequently treated with various combinations of vehicle, ER or PR-targeting drugs. Post treatments, xenografts were harvested and RNA-seq was performed. B. Cell viability of T47D cells in response to treatments with various combinations of PR agonist R5020, pure PR antagonist EC317 and selective PR modulator (SPRM) EC313. These drugs were treated at various concentrations (1 pM to 10 nM) mentioned on the horizontal axis. Vertical axis represents the cell numbers after the end of six days of treatments of interest. C., D. Heatmaps display unsupervised clustering of C. sample-sample correlations and D. gene expression observed in T47D xenografts treated with various combinations of ER and PR-targeting drugs. E. Immunoblots to measure ER and PR levels in T47D cells used to seed T47D xenografts. F. Immunoblots to measure ER and PR levels in various T47D xenografts used in the study. The immunoblots for individual and combination (with tamoxifen) therapies with CDB4124 and CDB4453 could not be included because of the lack of the starting material. G. Unsupervised clustering of sample-sample correlations observed between transcriptomes of T47D xenografts treated with vehicle, tamoxifen, SPRM EC313 alone or in combination with SERMs tamoxifen, bazedoxifene, raloxifene or selective ER-degrader fulvestrant. High correlation (i.e., correlation coefficient 1) between any two samples is displayed in red and low correlation (i.e., correlation coefficient 0) is displayed in blue.

Figure 5. Combination therapies with tamoxifen and SPRMs result in tumor regression (A - B)

Overall survival in METABRIC's discovery cohort as classified by high or low expression of A. PRA or B. PRB-regulated genes. Top 1000 differential PRA and PRB-regulated genes were obtained from Figure 1D and are provided in Supplementary Table 3 C. - D. Overall survival in METABRIC's discovery cohort as classified by high or low expression of top differential. C. PR antagonists or D. PR agonists-regulated genes. Top 100 PR agonist and antagonist-regulated genes were obtained from Figure 3C and are provided in Supplementary Table 4. E. T47D xenografts were grown in ovariectomized nude mice containing estrogen silastic implants and were treated with placebo, tamoxifen, CDB4453, EC313 or tamoxifen plus CDB4453 or EC313. Average tumor volume at the start of therapies was 125 mm³ and percentage change in tumor volume is shown (n > 16). F. ER+/PR- patient-derived xenografts were treated with placebo, tamoxifen, CDB4124 or tamoxifen plus CDB4124. Average tumor volume at the start of therapies was 125 mm³ and the total tumor volume is shown (n > 9). Mean and S.E.M are plotted. (^* <0.05 and ns not significant). G. In ER/PR crosstalk, distinguishing overlapping effects observed due to concomitant interaction of another receptor with its ligand (agonist or antagonist) from unique isoform and ligand-specific effects would guide the development of biomarkers for patient selection and translation of PR-targeted therapies to the clinic.

Figure 4. PR isoforms appropriate similar motifs at distinct locations to differentially reprogram estrogen signaling

A. - B. Heatmaps display intensity of sequencing obtained on anti-ER ChIP before (10 nM estrogen (E2) alone) and after (10 nM estrogen plus 10 nM R5020) remodeling by PRA A. or PRB B.. C. - D. Expression of estrogen and progestin-regulated genes in T47D cells expressing either PRA C. or PRB D.. Heatmaps are row-normalized and include the union of estrogen and progestin-regulated genes. E. Cellular pathways enriched in the genes that are differentially expressed in response to 10 nM R5020 treatment of T47D cells expressing PRA/PRB mixtures [1] or PRA and PRB individually. F. - G. Changes in F. matrigel invasion and G. cell confluence of ER+/PRA+ and ER+/PRB+ T47D cells in response to treatments with 10 nM estrogen or both 10 nM estrogen plus 10 nM R5020. ^* denotes P-value < 10^-1, ^*** P-value < 10^-3. Three biological replicates were performed and each of the experimental conditions had at least 12 technical replicates.