Progesterone inhibits epithelial-to-mesenchymal transition in endometrial cancer

Abstract

Background: Every year approximately 74,000 women die of endometrial cancer, mainly due to recurrent or metastatic disease. The presence of tumor infiltrating lymphocytes (TILs) as well as progesterone receptor (PR) positivity has been correlated with improved prognosis. This study describes two mechanisms by which progesterone inhibits metastatic spread of endometrial cancer: by stimulating T-cell infiltration and by inhibiting epithelial-to-mesenchymal cell transition (EMT).

Methodology and principal findings: Paraffin sections from patients with (n = 9) or without (n = 9) progressive endometrial cancer (recurrent or metastatic disease) were assessed for the presence of CD4+ (helper), CD8+ (cytotoxic) and Foxp3+ (regulatory) T-lymphocytes and PR expression. Progressive disease was observed to be associated with significant loss of TILs and loss of PR expression. Frozen tumor samples, used for genome-wide expression analysis, showed significant regulation of pathways involved in immunesurveillance, EMT and metastasis. For a number of genes, such as CXCL14, DKK1, DKK4, PEG10 and WIF1, quantitive RT-PCR was performed to verify up- or downregulation in progressive disease. To corroborate the role of progesterone in regulating invasion, Ishikawa (IK) endometrial cancer cell lines stably transfected with PRA (IKPRA), PRB (IKPRB) and PRA+PRB (IKPRAB) were cultured in presence/absence of progesterone (MPA) and used for genome-wide expression analysis, Boyden- and wound healing migration assays, and IHC for known EMT markers. IKPRB and IKPRAB cell lines showed MPA induced inhibition of migration and loss of the mesenchymal marker vimentin at the invasive front of the wound healing assay. Furthermore, pathway analysis of significantly MPA regulated genes showed significant down regulation of important pathways involved in EMT, immunesuppression and metastasis: such as IL6-, TGF-β and Wnt/β-catenin signaling.

Conclusion: Intact progesterone signaling in non-progressive endometrial cancer seems to be an important factor stimulating immunosurveilance and inhibiting transition from an epithelial to a more mesenchymal, more invasive phenotype.

Figure 1. Expression and histological distribution of PRA+PRB and CD4+, CD8+ and Foxp3+ T-lymphocytes in primary endometrial carcinoma specimens.

A: Overview of immunohistochemical staining for CD4, CD8 and FOXP3 in primary endometrial cancer specimens in non-progressive disease (n = 9) compared to progressive disease (n = 9) (magnification 0,4x, inlay 10x). Non-progressive disease shows pronounced staining, whereas progressive disease shows reduced staining. The scale-bar represents 10 mm. B: Quantification of CD4, CD8 and FOXP3 cell counts on the tumor edge (Tumor Edge), in the tumor (Intratumoral) and on the endometrial-myometrial border (EM border). *indicates a p-value<0.05 (Mann-Whitney U-test). C and D: Representative non-progressive (C) and progressive (D) patient tissues were stained for CD4, CD8 and PRA+PRB and show a positive correlation between the presence of TILs and the expression of PR. Magnification is 5x and the scale-bar represents 1 mm. Patients 6 and 11 were both included in the micro-array analyses. Furthermore patient 11 had only recurrent disease, while patient 12 had recurrent and metastatic disease.

Figure 2. RT-PCR results of genes of interest in the patient samples.

CXCL14, DKK1, DKK4, WIF1 and PEG10 were selected from the micro-array results and verified with real time RT-PCR. Significance was calculated using a Mann-Whitney U-test. A p-value of 0.05 was considered to be statistically significant.

Figure 3. Progesterone induced inhibition of migration in a wound-healing assay.

IKPRA-1 (A), IKPRB-1 (B) and IKPRAB-36 (C) cells were cultured in the absence (white bullets) or presence (black bullets) of 1 nM MPA and used for a wound-healing assay (n = 3) and closure of the wound was measured as a percentage of total closure (100% means the wound is open, 0% means the wound has closed). D shows representative images of the process of wound-healing with in red the wound. E shows IF for nuclei (DAPI) and vimentin expression on the invasive front of the manually inflicted wound. In this figure, the wound was always situated on the right side.

Figure 4. Invasion of PR positive Ishikawa EC cell lines.

IKPRA-1, IKPRB-1 and IKPRAB-36 cells were cultured in the absence (black dots) or presence (white dots) of 1 nM MPA in a modified Boyden chamber. After 96 hours, cells that had migrated through the pores of the upper well were counted. The figure represents three independent experiments performed in triplicate. *indicates a p-value of <0.05 (Mann-Whitney U-test).

Figure 5. MPA induced regulation of TGF-β and Wnt/β-catenin signaling in the IKPRAB-36 cell line.

A and B: In these pathways a green color represents down regulation by MPA and a red color represents up regulation by MPA. Signaling pathways were provided by Ingenuity Pathway Assist Software© and individual gene expression levels are available in Table S4. C: Western blot showing FOXO1 expression in the IKPRA-1, IKPRB-1, IKPRAB-36 and IKLV-8 cell lines cultured in the absence (control) or presence (MPA) of 1 nM MPA. *indicates significant regulation in the micro-array analysis (Table S4).