Identification of miRNAs differentially expressed in the Sca-1 subpopulations of the COMMA-DβGeo cell line. (A) Representative FACS histogram showing Sca-1 expression in COMMA-DβGeo cells. Cells stained with a FITC-conjugated antibody against Sca-1 analyzed by FACS display two distinct populations based on Sca-1—FITC fluorescence: 17% are Sca-1⁺ and 60.2% are Sca-1⁻. Shown are 75,000 events. (B) Representative heat map showing the differential expression of miRNAs in the Sca-1⁺ and Sca-1⁻ subpopulations. Each row represents a different mature miRNA sequence. Shown are median values from normalized, log-ratio (base 2) data sets and plotted as a heat map from two biological and technical replicates. Yellow indicates high miRNA expression; blue indicates low miRNA expression. Differentially expressed miRNAs are detailed in Table 1 from three independent microarray experiments performed on three biological replicates each. Expression values produced a cladogram reflecting a clear distinction between the Sca-1⁺ and Sca-1⁻ cells. The red arrow indicates the location of miR-205 on the array, where the average fold change was 6.56 higher in the Sca-1⁺ cells versus Sca-1⁻. (C) TaqMan qPCR of miR-205 verify the microarray data. miR-205 is approximately 19-fold higher in the Sca-1⁺ population compared with the Sca-1⁻ population. Data represent the mean expression levels of two different biological replicates calculated by ΔΔC_T normalized to the endogenous snoRNA55 control, of two independent experiments performed in triplicate. ***P<0.0001.

Overexpression of miR-205 increases the colony-forming capacity of Sca-1⁻ COMMA-βGeo cells. COMMA-βGeo cells transduced with the pSM2 miRNA overexpression vectors were sorted into Sca-1⁺ and Sca-1⁻ populations. (A) Left y-axis: fold change of Sca-1⁺ colonies compared with the luciferase shRNA control. Right y-axis: number of colonies formed. When overexpressing miR-205, the clonogenic potential of the Sca-1⁺ COMMA-βGeo significantly increases. **P<0.0001. (B) Left y-axis: fold change of Sca-1⁻ colonies compared with the luciferase shRNA control. Right y-axis: average number of colonies formed. When Sca-1⁻ cells overexpress miR-205 they gain significant potential to form colonies when normally they could not. **P<0.0001. (C) Representative images of colonies formed on Matrigel when sorted for Sca-1. Scale bars: 100 μm. Arrows highlight colonies formed. (D) Average area of colonies formed by Sca-1⁺ cells. Colonies were measured using Image-Pro Plus (Media Cybernetics). The average colony area formed from Sca-1⁺ cells overexpressing miR-205 were statistically larger than those formed from the luciferase shRNA control cells. An average of 130 colonies measured from each group were measured. ***P<0.0005. Data represent the mean of three independent experiments with two biological replicates.

Targets of miR-205. (A,B,G-I) SYBR Green qPCR of putative targets of miR-205; (C-E) TaqMan qPCR of putative targets of miR-205. Data represent the mean expression levels of two biological replicates calculated by ΔΔC_T normalized to either endogenous β-actin (A,B,G-I) or 18S rRNA (C-E), of two independent experiments performed in triplicate. (A-C) Atp1a1, Dok4 and ZEB2 are transcripts identified by microarray analysis that decreased when miR-205 was overexpressed. These transcripts were also predicted targets of miR-205 by several algorithms (Table 2). (A) Atp1a (ATPase, Na⁺-K⁺ transporting, α1 polypeptide) expression is 1.78-fold lower in miR-205-overexpressing cells (**P=0.005) and is a predicted miR-205 target by miRNAMap and miRNA Viewer. (B) Dok4 expression is 2.28-fold lower in miR-205-overexpressing cells (*P=0.007) and is a predicted miR-205 target by miRNA Viewer, TargetRank and TargetScan. (C) ZEB2 expression is 2.41-fold lower in miR-205-overexpressing cells (*P=0.005) and is a predicted miR-205 target by miRNAMap, TargetRank and TargetScan. (D) ZEB1 expression is 1.7-fold in miR-205-overexpressing cells (*P<0.05) and is a predicted miR-205 target. (E) Expression of E-cadherin increases 2.3-fold as a result of ZEB1 and ZEB2 decreasing following miR-205 overexpression (***P=0.0008). (F) Stably transduced COMMA-βGeo cells were stained for E-caherin (green) and vimentin (red). As a result of lower ZEB1 and ZEB2 expression following miR-205 overexpression, membrane-bound E-cadherin is upregulated. Some membrane-bound E-cadherin can be detected in cells expressing the luciferase shRNA, but only at areas of cell-cell junctions (inset). Scale bars: 25 μm. (G-I) Six5, Ppp2r4 and Tcp1 expression levels decrease when miR-205 is overexpressed; however they are not predicted miR-205 targets by any of the available algorithms. (G) Six5 (Six5 sine oculis-related homeobox 5 homolog) expression is 1.94-fold lower in miR-205-overexpressing cells (***P=0.009). (H) Ppp2r4 (protein phosphatase 2A, regulatory subunit B) expression is 1.46-fold lower in miR-205-overexpressing cells (*P=0.004). (I) Tcp1 (T-complex protein 1) expression is 1.64-fold lower in miR-205-overexpressing cells (*P=0.002).

Overexpression of miR-205 affects cell size, cell proliferation and expands the progenitor-cell compartment. (A) Representative FACS analysis of COMMA-DβGeo cells indicates that overexpression of miR-205 expands the Sca-1⁺ percentage compared with cells expressing the luciferase shRNA control. Cells were sorted for Sca-1—FITC and separated into Sca-1⁺ and Sca-1⁻ cells. Shown are 25,000 events. (B) Quantitation of the percent of cells that are either Sca-1⁺ or Sca-1⁻ based on FACS analysis. Populations that overexpress miR-205 (white bars) have more Sca-1⁺ cells compared with populations expressing the luciferase shRNA control (black bars). ***P<0.001; **P<0.01. (C) COMMA-DβGeo cells overexpressing miR-205 are smaller than cells expressing the control luciferase shRNA as determined by FACS analysis. Cells were sorted for Sca-1—FITC and separated into Sca-1⁺ and Sca-1⁻ cells. In all three compartments — total (viable) cells, Sca-1⁺ and Sca-1⁻ — miR-205-overexpressing cells are all smaller than the luciferase shRNA cells. Shown are 25,000 events. (D) Quantitation of mean FSC percent, which is proportional to cell size, based on FACS analysis. miR-205-overexpressing cells (white bars) are significantly smaller in the Sca-1⁺ and Sca-1⁻ cell populations compared with cells expressing the luciferase shRNA control (black bars). Cell size is given in percentage relative to the control, set at 100%. *P<0.05; **P<0.005. (E) Growth curves of miR-205-overexpressing cells and luciferase shRNA cells. Cells were plated at the same density and counted every 24 hours for 6 days (shown are timepoints up to 96 hours). Cells overexpressing miR-205 are proliferating at significantly higher rates than the control. At 144 hours post-plating, there are still significantly more miR-205-overexpressing cells (P<0.0001) compared with the luciferase shRNA cells (not shown). ***P<0.0001 compared with luciferase shRNA. (F) Cell-proliferation assay, where percent dividing cells represents the fraction of cells that are or have synthesized DNA in preparation for cell division (equivalent to the sum of the percentages of cells in S+G₂-M), as determined by the Click-iT EdU Flow Cytometry Assay. Cells overexpressing miR-205 have significantly higher proliferation rates (P<0.005) compared with the luciferase shRNA cells. All data represent the mean of two independent experiments performed in duplicate with two biological replicates.

miR-205 targets the mRNA of the tumor-suppressor protein PTEN. (A) Representation of the pGL3-PTEN 3′UTR reporter plasmid. Shown are the computationally predicted binding sites for mmu-miR-205 in the PTEN 3′UTR (red arrowheads). Mutations were made to the seed region of the miR-205-binding sites to disrupt miRNA-target binding (mutated nucleotides shown in red). (B) Luciferase reporter activity of the PTEN 3′UTR in NIH/3T3 cells expressing miR-205, the non-silencing hairpin negative control or the luciferase shRNA positive control. Reporter activity decreases by 30% when cells are overexpressing miR-205. Each sample was normalized to Renilla luciferase activity. *P<0.01 or **P<0.005 relative to the non-silencing hairpin. RLU, relative luciferase units (firefly luciferase/Renilla luciferase). (C) PTEN 3′UTR reporter activity of the in NIH/3T3 cells expressing the miR-205Zipper, the non-silencing hairpin negative control or the luiferaseZipper control. When expressing the miR-205Zipper, reporter activity increases 2.5-fold relative to the non-silencing control. Each sample was normalized to Renilla luciferase activity. ***P<0.0005 relative to the non-silencing hairpin. (D) The computationally predicted miR-205-binding sites in the PTEN 3′UTR were mutated in the seed region. Single seed-region mutations in the 559 and 732 sites significantly relieved silencing of the reporter. *P=0.013 or **P=0.0057. Results are representative of four individual experiments with two biological replicates for each overexpression cell line. (E, left) Western blot analysis of PTEN, phospho-GSK-3β (Ser9) and total GSK-3β levels in COMMA-DβGeo cells transduced with pSM2 miRNA overexpression vectors. When overexpressing either miR-205 or the PTEN shRNA, PTEN expression levels decrease compared to luciferase shRNA control. The blot was stripped and reprobed with β-actin as a loading control. (E, right) Levels of phospho-GSK-3β increase as a result of lowered PTEN levels when cells are overexpressing miR-205, compared with the luciferase shRNA control. The blot was stripped and reprobed with GAPDH as a loading control. Quantitation of relative expression levels of PTEN to β-actin and phospho-GSK-3β to total GSK-3β were determined with ImageJ. (F) TaqMan qPCR of PTEN expression. PTEN expression decreases approximately twofold when miR-205 is overexpressed, as well as when the PTEN shRNA is expressed, relative to the luciferase shRNA control. Data represent the mean expression levels of two biological replicates calculated by ΔΔC_T normalized to endogenous 18S rRNA. ***P<0.0001.

miR-205 is highly expressed in mouse mammary-gland stem-cell populations. Primary mammary epithelial cells were isolated from wild-type Balb/c virgin females, age 8-9 weeks, and separated by FACS for cell-surface markers and analyzed by TaqMan qPCR for miR-205 expression. Data represent the mean expression levels of two different biological replicates calculated by ΔΔC_T normalized to the endogenous snoRNA55 control, of two independent experiments performed in triplicate. (A) Representative FACS profile for CD24-FITC and Sca-1—PE. (B) Cells from different CD24/Sca-1 subpopulations were collected and assayed for expression of miR-205. In the basal and myoepithelial Lin⁻/CD24^+/lo/Sca-1⁻ stem-cell-enriched population, miR-205 expression is approximately 25-fold higher than in the lumenal Lin⁻/CD24^+/hi/Sca-1⁺ and Lin⁻/CD24^+/hi/Sca-1⁻ populations. (C) Representative FACS profile for CD29-FITC and CD24-PE. (D) Cells from different CD29/CD24 subpopulations were collected and assayed for expression of miR-205. In the Lin⁻/CD29^hi/CD24⁺ self-renewing stem-cell-enriched population, miR-205 expression is approximately 85-fold higher than in the non-stem-cell populations, including the lumenal progenitor Lin⁻/CD29^lo/CD24⁺ population. (E) Representative FACS profile for CD24-PE and CD49f-FITC. (F) Cells from different CD24/CD49f subpopulations were collected and assayed for expression of miR-205. In the stem-cell-enriched MRU population (Lin⁻/CD24^med/CD49f^hi), miR-205 expression is approximately 45-fold higher, and the myoepithelial MYO population (Lin⁻/CD24^lo/CD49f^lo) is approximately 32-fold higher, than in the non-stem-cell Lin⁻/CD24^+/hi/CD49f⁻ population.