Estrogen receptor beta reduces colon cancer metastasis through a novel miR-205 - PROX1 mechanism

Abstract

Colon cancer is a common cause of cancer death in the Western world. Accumulating evidence supports a protective role of estrogen via estrogen receptor beta (ERβ) but the mechanism of action is not known. Here, we elucidate a molecular mechanism whereby ERβ represses the oncogenic prospero homebox 1 (PROX1) through the upregulation of miR-205. We show that PROX1 is a potential target of miR-205 and that in clinical specimens from The Cancer Genome Atlas data, ERβ and miR-205 are decreased in colorectal cancer tissue compared to non-tumorous colon, while PROX1 levels are increased. Through mechanistic studies in multiple colorectal cancer cell lines, we show that ERβ upregulates miR-205, and that miR-205 targets and represses PROX1 through direct interaction with its 3'UTR. Through the generation of intestine-specific ERβ knockout mice, we establish that this pathway is correspondingly regulated in normal intestinal epithelial cells in vivo. Functionally, we demonstrate that miR-205 decreases cell proliferation and decreases migratory and invasive potential of colon cancer cells, leading to a reduction of micrometastasis in vivo. In conclusion, ERβ in both normal and cancerous colon epithelial cells upregulates miRNA-205, which subsequently reduces PROX1 through direct interaction with its 3'UTR. This results in reduced proliferative and metastatic potential of the cells. Our study proposes a novel pathway that may be exploited using ERβ-selective agonists and/or miR-205-replacement therapy in order to improve preventive and therapeutic approaches against colon cancer.

Keywords: PROX1; colorectal cancer; estrogen receptor; metastasis; microRNA.

Figure 1. Expression of ERβ, miR-205, and PROX1 in human colon tissues and cells

(A) In primary colorectal tumor samples ERβ mRNA levels are decreased compared to non-tumorous tissue. This is accompanied by decreased miR-205 and increased PROX1. Data were collected from The Cancer Genome Atlas (TCGA), and illustrated using box plots. RSEM was used for transcript quantification of mRNAs. RPM indicates reads per million, for miRNAs. (P < 0.05, Student's t-test) (B) Spearman correlation between ERβ (ESR2) and PROX1 mRNA expression in corresponding TCGA colon tissue. (C) Levels of miR-205 is inversely related to PROX1 protein in human colon cancer cell lines SW480, HT29, HCT116, SW403, and SW620. Relative miR-205 levels were determined using qPCR and PROX1 protein levels using western blot. PROX1 protein is low in HT29, but clearly visible when using a longer exposure time (see Figure 3B).

Figure 2. ERβ upregulates miR-205 expression in colon cancer cells

(A) Western blotting demonstrates FLAG-tagged ERβ protein levels after lentivirus-mediated transduction in SW480, HT29, and HCT116 cells, compared to non-detectable levels in control cells. Recombinant ERβ (59 kDa) was used as positive control (left lane), and β-actin as loading control. (B) Mature miR-205 expression is upregulated by ERβ and is dependent on its DNA-binding domain. SW403 and SW620 were transiently transfected by 500 ng pcDNA3.1, ERβ or ERβ-mDBD, and expression determined using miRNA qPCR and normalized to U6 snRNA 48 h after transfection. Transfection and subsequent analysis were replicated three times. (C) ERβ regulates primary miR-205 in SW480 cells. Expression normalized to 18S. (*P < 0.05, **P < 0.01, Student's t-test).

Figure 3. miR-205 directly silences PROX1 by targeting its 3′UTR

(A) PROX1 mRNA levels are downregulated by miR-205 overexpression in colon cancer. SW480-ERβ and HT29-ERβ cells were transfected with 50 nM of miR-205 mimic or scrambled mimic control in three replicates, followed by qPCR analysis 48 h after transfection. (B) PROX1 protein is repressed by miR-205. Protein was extracted 72 h after single miR-205 mimic or scrambled mimic control transfection and β-actin was used as loading control. (C) miR-205 inhibitor upregulates PROX1 mRNA levels in colon cancer cells. SW480-ERβ and HT116-ERβ cells were transiently transfected by miR-205 inhibitors or inhibitor control at the final concentration of 50 nM, followed by qPCR analysis after 48 h. Experiment replicated two times. (D) Sequence alignment between human PROX1 3′UTR and mature miR-205 (miRanda). The position refers to distance from the start of 3′UTR. Highlighted nucleotides indicate the seed sequence of miR-205. The mutant human PROX1 3′UTR is represented in the lower panel and the disruption of base-pairing is indicated by X. (E) miR-205 directly interacts with the 3′UTR of PROX1. HEK293 cells were co-transfected with wild-type or mutant PROX1 3′UTR luciferase construct (800 ng), and miR-205 mimic or scrambled mimic control (50 nM). Luciferase activity was normalized to Renilla luciferase 24 h after transfection and depicted as the mean ± S.D. The experiment was replicated three times. (*P < 0.05, **P < 0.01, ***P < 0.001, Student's t-test).

Figure 4. Intestinal-specific KO of ERβ results in decreased miR-205 in colon epithelial cells

Colon scraping samples were collected from wild-type mice (N = 15) and ERβ^iKO (N = 16) and RNA expression was analyzed by qPCR. (*P < 0.05, Student's t-test). Supplementary Table S1 shows the corresponding correlation between ERβ, miR-205, and PROX1 expression.

Figure 5. Both ERβ and miR-205 block cell proliferation

(A) Cell proliferation is inhibited by ERβ expression in SW480 and HT29 cells. BrdU measurment was performed 60 min after BrdU was added to cells engineered to express ERβ and corresponding controls. (B) Cell proliferation is repressed by miR-205 transfection. SW480 and HT29 cells were transfected by miR-205 mimic or scrambled control (50 nM, single transfection), followed by BrdU measurment 48 h after transfection. (C) Distribution of cell cycle is similarly affected by ERβ and miR-205 overexpression in colon cancer cell lines. Flow cytometry was performed after propidium iodide (PI) staining to assess cell cycle distribution. All experiments were replicated three times. NC: scrambled negative control. (*P < 0.05 compared to control, Student's t-test).

Figure 6. ERβ and miR-205 affects EMT, stemness and cell adhesion in colon cancer

(A) EMT markers and PROX1-regulated genes FN1 and SNAIL are regulated by both ERβ and miR-205 in SW480 and HT29 cells. mRNA levels were analyzed 48 h after miR-205 double transfection, and experiments were replicated two times. (B) Cell adhesion genes are regulated by ERβ in SW403 cells. mRNA levels were analyzed 72 h after ERβ transfection, and experiments were replicated three times. (C–D) Colon cancer stemness is affected by ERβ, miR-205, and PROX1 in SW480 and/or HT29 cells. Cells were analyzed 48 h after single miR-205 mimic and siR-PROX1 transfection, and experiments were replicated two times. (E) ERβ and miR-205 regulate cellular adhesion in several colon cancer cell lines. Cell adhesion assay was performed 72 h after ERβ or miR-205 mimic transfection, and all experiments were replicated three times. Absorbance (Abs) indicates the relative amount of cells that attach to the collagen surface, and is a measure of adhesion. NC: scrambled negative control mimic. (*P < 0.05, **P < 0.01, ***P < 0.001, Student's t-test).

Figure 7. ERβ and miR-205 inhibits tumor invasion in vitro and in vivo

(A) Cell invasiveness is reduced by ERβ and miR-205. SW480 and HT29 cells were transfected and transwell assay performed. After 24 h starvation of cells, 10% FBS was used as chemoattractant at the bottom of the chamber. Cells were allowed to migrate for 12 h prior to staining and quantification. (*P < 0.05, **P < 0.01, Student's t-test) (B–C) In vivo metastasis assay. Labeled SW480 and HT29 cells, with and without ERβ expression, or with and without transfection of miR-205 mimic, were injected into transgenic Tg(kdrl:EGFP)mitfab692 zebrafish larvae in three independent experiments and observed for metastatic potential after 24 and 48 hpi. Both ERβ and miR-205 exhibited a significant (p < 0.05) anti-metastatic potential in both cell lines.