Sequence variation among members of the miR-200 microRNA family is correlated with variation in the ability to induce hallmarks of mesenchymal-epithelial transition in ovarian cancer cells

HEY and HEY A8 ovarian cancer cell lines were provided by Gordon B. Mills (MD Anderson Cancer Center, Houston, TX). SKOV-3 ovarian cancer cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). Cells were cultured in RPMI 1640 (Mediatech, Manassas, VA) supplemented with 10% FBS (Fetal Bovine Serum; Atlanta Biologicals, Lawrenceville, GA) and 1% antibiotic-antimycotic solution (Mediatech-Cellgro, Manassas, VA). For miRNA transfections, 6 × 10⁴ cells were seeded per well in 24-well plates. Cells at exponential phase of growth were transfected with 30 nM miRNA purchased as Pre-miR miRNA Precursors (Ambion, Austin, TX) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) and according to the manufacturer’s instructions. Cells were allowed to grow for 48 hours before RNA isolation. Ambion Pre-miR miRNA Precursor Negative Control was used as a negative control (nc-miR).

Morphological changes were monitored using an Olympus IX51 microscope (Olympus Optical, Melville, NY). The effect of treatment on cell morphology was objectively measured using CellProfiler cell-imaging software (2.1.0)[11].

The primer sequences employed are as follows: KRT7 (keratin 7): forward 5′- GGACATCGAGATCGCCACCT-3′ and reverse 5′- ACCGCCACTGCTACTGCCA-3′; KRT8 (keratin 8): forward 5′- CCGTGGTTGTGAAGAAGATCG-3′ and reverse 5′-GCTGTTCACTTGGGCAGGAC-3′; KRT18 (keratin 18): forward 5′-TGAGACGTACAGTCCAGTCCTT-3′ and reverse 5′-GCTCCATCTGTAGGGCGTAG-3′; GAPDH: forward 5′- TGCACCACCAACTGCTTAGC -3′ and reverse 5′- GGCATGGACTGTGGTCATGA -3′. The primer sequences for mesenchymal biomarker genes were described previously[8]. The threshold cycle and ΔΔCt method was used for calculating the relative amount of the target RNA. Expression values were normalized using GAPDH as a reference gene.

Aliquots of cells were seeded in 96-well plates and treated with nine concentrations of cisplatin ranging from 0.1-50 μM. After 72 hours, TOX-8 reagent (Resazurin based in vitro toxicology assay kit, Sigma-Aldrich, St Louis, MO) was added to the wells and fluorescence was measured (λex = 560 nm, λem = 590 nm). Outlier values were removed using Grubbs’ test. The ratio of the background-subtracted fluorescence intensities of drug treated to untreated cultures was calculated in percentages across all concentrations. IC₅₀ values were determined by non-linear regression of log-transformed data using a normalized response-variable slope model with GraphPad Prism v.6 (GraphPad Software Inc., La Jolla, CA).

Statistical significance of differences in qRT-PCR experiments between experimental and control samples was determined using a two-tailed Student’s t-test. Significance of differences in eccentricity profiles of experimental and control samples was evaluated using Mann-Whitney U test. Statistical significance of differences in mean IC₅₀ values among miRNA transfections was tested using analysis of variance (ANOVA) and Tukey’s multiple comparison test as a post-hoc test. All experiments were performed using at least three biological replicates.

We have previously shown that exogenous over expression of miR-429 induces MET in a well-characterized mesenchymal-like OC cell line (HEY)[8]. We were interested in determining if this ability of miR-429 to induce MET extends to other mesenchymal-like OC cell lines as well. Consistent with our previous results, exogenous over expression of miR-429 in two additional mesenchymal-like OC cell lines (HEY A8[12] and SKOV-3[13]) was found to induce morphological and molecular changes consistent with MET (Figure 1). Both SKOV-3 and HEY A8 cells display a change from the elongated shape characteristic of mesenchymal cells to the more cuboidal shape characteristic of epithelial cells after transfection with miR-429. This morphological change was objectively validated using the CellProfiler cell image analysis software[11]. Changes in expression of two epithelial KRT8 and KRT18 and two mesenchymal [ZEB1 and ZEB2 (zinc finger E-box binding homeobox 1 and 2)] molecular biomarkers were consistent with MET.

MiR-429 is a member of the miR-200 family of miRNAs (Figure 2A). All members of the miR-200 family have previously been either directly or indirectly implicated in EMT[9, 10]. Having established that ectopic over expression of miR-429 induces MET in a variety of mesenchymal-like OC cell lines, we were interested in determining if this MET-inducing potential extends to other members of the miR-200 family as well.

The results presented in Figure 2 demonstrate that over expression of all miR-200 family members results in a significant change from the elongated, spindle-shaped morphology of the mesenchymal-like cells to the more rounded, cuboidal morphology characteristic of epithelial cells. No detectable change in morphology was observed in cells treated with the negative control (nc-miR).

The expression levels (qRT-PCR) of a series of previously established epithelial and mesenchymal biomarkers[14] were monitored in HEY cells transfected with members of the miR-200 family. Consistent with acquisition of a more epithelial phenotype, expression levels of all of the epithelial biomarkers (KRT8, KRT18, KRT7) were increased after over expression of each member of the miR-200 family (Figure 3). As expected, the mesenchymal biomarkers ZEB1/ZEB2 displayed a general reduction in expression levels after over expression of each member of the miR-200 family. The mesenchymal biomarker FN1 was also significantly down regulated in cells transfected with miR200b, miR200c and miR429.

Human miRNAs are known to be capable of regulating expression of their target genes by modulating mRNA levels and/or blocking translation[15]. Thus, to further explore the inconsistent response of changes in FN1 RNA levels after transfection by individual members of miR200 family, we additionally monitored expression changes on the protein level by immunofluorescence staining. The epithelial biomarker, CDH1[14, 16], was included in these assays as an additional control (Figure 4).

Consistent with the observed increase in expression levels of each of the epithelial biomarkers monitored on the RNA level, CDH1 protein was detected in all transfected cells but was absent in un-transfected cells and in cells transfected with the negative control. Interestingly, levels of FN1 protein were generally consistent with what was observed on the RNA level, i.e., protein levels were significantly down regulated in cells transfected by miR-200b, miR-200c or miR-429 but unchanged or slightly up-regulated in cells transfected by miR-200a, miR-141 (Figure 4). Collectively, these findings suggest that FN1 is a regulatory target of miR-200b, miR-200c and miR-429 but is not subject to regulation by miR-200a and miR-141. In addition, since a change from a mesenchymal to an epithelial phenotype was observed in cells transfected by all members of the miR-200 family, we conclude that FN1 is non-essential for MET in OC.

Previous studies indicate that the most functionally significant component of miRNA/mRNA pairing involves the miRNA “seed region”, i.e., nucleotides 2-8 on the 5′ end of the miRNA[17]. Although all members of the miR-200 family are sequentially distinct, miR-200b, miR-200c and miR-429 are sequentially identical within their respective seed regions (Figure 2A). This identity may explain their shared ability to down regulate FN1 when ectopically over expressed in HEY cells. To further explore this possibility, we utilized three miRNA target prediction algorithms (miRanda[18], miRDB[19], and TargetScan[20]) to identify members of the miR-200 family that are predicted to directly target FN1 mRNA and the other biomarkers monitored in our PCR experiments. The results, presented in Table 1, indicate that all three algorithms predict the presence of binding sites in FN1 mRNA for miR-200b, miR-200c and miR-429, but not for miR-200a, miR-141. These computational predictions are consistent with our experimental results and indicate that differences in target specificity for FN1 mRNA between the miR-200 family members accounts for the observed variation in regulation of FN1 expression.

Previous studies indicate that EMT is often associated with decreased sensitivity of a variety of cancer cells to chemotherapy[21‐23]. Consistent with this observation, OC cells undergoing EMT have been reported to display a decreased sensitivity to platinum-based drugs, a frequently employed first line therapeutic in the treatment of OC[24, 25]. Since EMT is reported to decrease the sensitivity of OC epithelial cells to platinum-based drugs, we were interested in determining if miR-induced MET of mesenchymal-like OC cells would induce an opposite effect, i.e., be associated with an increased sensitivity of OC cells to platinum-based drugs. To explore this possibility, we monitored the drug susceptibility (IC₅₀) of HEY cells to cisplatin after miRNA-induced MET relative to controls.

The results presented in Figure 5 demonstrate that ectopic over expression of all members of the miR-200 family results in a significant increase in cisplatin sensitivity relative to negative control. While notable variation was observed among cells transfected by different members of the miR-200 family, in most cases this variation was not statistically significant. Cells transfected with miR-200b were significantly more sensitive to cisplatin than those transfected with miR-429 (p < 0.05). Interestingly, these two miRNAs share identical seed regions suggesting that non-seed sequences may be of significance in miR-200b induced drug sensitivity.