Regulation of miR-200 family microRNAs and ZEB transcription factors in ovarian cancer: evidence supporting a mesothelial-to-epithelial transition

Abstract

Objective: Our objective was to characterize the expression and function of the miR-200 family of microRNAs (miRNA) in ovarian carcinogenesis.

Methods: We used qRT-PCR to examine expression of the miR-200 miRNA family and its predicted targets, the ZEB1 and ZEB2 transcriptional repressors, in primary cultures of normal cells from the surface of the ovary and in a panel of 70 ovarian cancer tissues and 15 ovarian cancer cell lines. We studied the mechanisms of regulation of miR-200 miRNAs and ZEB transcription factors in ovarian cells using 3' UTR luciferase reporters, promoter luciferase reporters and siRNAs.

Results: miR-200 family members are expressed at low or negligible levels in normal ovarian surface cells and substantially increase in expression in ovarian cancer, whereas expression of ZEB1 and ZEB2 shows the opposite pattern. There is reciprocal repression between miR-200 family members and ZEB transcription factors, creating a double negative regulatory feedback loop resembling that reported in other cancer cell types. In contrast to epithelial cells from other sites, expression levels of miR-200 miRNAs and ZEB1/2 in cells from the ovarian surface are more consistent with a mesenchymal cell phenotype, potentially reflecting the mesothelial origin of the ovarian surface.

Conclusion: Analysis of ovarian cancer tissues suggests that ovarian surface cells acquire a more epithelial miR-200-ZEB1/2 phenotype as they undergo transformation, switching from a miR-200 familyLOW and ZEB1/2HIGH state to a miR-200 familyHIGH and ZEB1/2LOW phenotype. Collectively, our data support the mesothelial-to-epithelial (Meso-E-T) model for development of ovarian cancers that arise from ovarian surface cells, as has been proposed previously on the basis of studies of protein markers.

Figure 1. Expression of miR-200 family members and ZEB1/2

(A) Real time qRT-PCR was performed across 70 primary tumor samples, three primary HOSE cell cultures and 15 different ovarian cancer cell lines for miR-200a, miR-200b, miR-200c, miR-141 and miR-429. Representative graphs for miR-200a and miR-200b are in the left column, additional graphs are in Figure S1. miR-200 family expression in HOSE and ovarian cancer cell lines are represented in log scale graphs in Figure S2. (B) The same samples were assayed for ZEB1 and ZEB2 expression (right column). Data are presented as fold change relative to the highest expression among the four HOSE samples.

Figure 2

Expression of transcription factor ZEB1 vs. copy number of miR-200a (left) and miR-200b (right) for each of the primary HOSE, primary ovarian tumor samples and ovarian cell lines (OvCa). Representative graphs are shown here; additional graphs for other miR-200 family members and ZEB2 are in Figure S3.

Figure 3. Regulation of ZEB2 by miR-200 family members

(A) Phase-contrast micrographs of 2008 and ES-2 cells, showing epithelial and mesenchymal morphologies, respectively. 2008 cells express high levels of miR-200 (left), whereas ES-2 cells do not express miR-200 (right). The asterisks denote cases in which expression values were below the limit of accurate linear quantitation. (B) 2008 (left) or ES-2 (right) cells were either mock transfected, or transfected with luciferase gene constructs containing either no ZEB2 3’ UTR (Control), wild-type ZEB2 3’UTR (wt) or ZEB2 3’ UTR with mutations at all conserved, predicted miR-200 binding sites (mut). Results from luciferase assays performed 24h after transfection are shown. (C) ES-2 cells were co-transfected with one of the three luciferase gene constructs described in (A) plus different quantities of pSM30-miR-200a or pSM30-miR-200b. 1/5x, 1x and 5x indicate the molar quantity of the pSM30-miR expression plasmid relative to the luciferase gene plasmid. Results show the mean and standard error of experiments carried out in triplicate. Firefly luciferase (FL) activity normalized to Renilla luciferase (RL) activity is shown on the y-axis.

Figure 4. ZEB1, ZEB2 transcription factors regulate miR-200c/141 cluster expression in ovarian cancer

(A) A fragment containing 922-bp of the miR-200c/141 promoter was cloned upstream of the luciferase gene of the pGL3-Enhancer vector. Single or multiple mutations were made in the conserved ZEB1/2 binding sites. Wild-type binding sites are shown in green, and mutated in black. (B) ES-2 cells with high levels of ZEB1 and ZEB2 were transfected with the pGL-3 Control (minimal 3’UTR), the wild-type or mutated miR-200c/141 promoter-luciferase constructs. Experiments were carried out in triplicate, and shown as the means and standard errors of Firefly luciferase activity normalized to Renilla luciferase activity. The experiment was repeated four times and the same expression patterns were seen. (C) Luciferase gene expression in ES-2 cells co-transfected with the wild-type miR-200c/141 promoter-luciferase construct and control, non-targeting siRNA (siNT) or siRNA against ZEB1 or ZEB2. (D) Real time qRT-PCR for ZEB1 and ZEB2 expression in samples transfected with control siRNA, or siZEB1 and siZEB2.

Figure 5. A model of epithelial ovarian carcinogenesis that includes mesothelial-to-epithelial transition

Our model supports the idea that ovarian surface mesothelial cells initially undergo a Mesothelial to Epithelial Transition (Meso-ET), acquiring characteristics of epithelial cells such as high miR-200, low ZEB1/2, and high E-cadherin levels (11, 31, 32). Subsequently during tumor progression, Epithelial to Mesenchymal Transition (EMT) may occur as has been described in other epithelial cancers (46, 47), based in part on a decrease in E-cadherin protein levels mediated by proteolytic cleavage (40). In further support of EMT at later stages of ovarian cancer progression is expression of ZEB2 in cells isolated from effusions (37-39); the expression of the miR-200 family miRNAs in ascites cells has not yet been examined.