Differential expression of miR-1, a putative tumor suppressing microRNA, in cancer resistant and cancer susceptible mice

RNA isolation

Animal studies were approved by the Ohio State University (OSU) and University of California, San Francisco Institutional Animal Care and Use Committees. The OSU IACUC determined that the research performed at OSU was exempt from IACUC review as pre-existing or commercially available animal specimens were used for this study. Fresh frozen skin samples from three FVB/NJ and three SPRET/EiJ female mice aged 4–5 weeks were obtained through the Jackson Laboratory surgical service. RNA was isolated from the skin using standard Trizol extractions (Invitrogen, Grand Island, NY). We isolated RNA a second time from two of the skin samples to generate a replicate control for the microarrays. Samples were DNAase treated. RNA was quantitated using Nanodrop (ThermoScientific, Wilmington, DE). RNA was isolated from eleven cSCCs from chemically treated C57Bl6/FVB F1 mice using standard Trizol methods. RNA samples were DNAase treated.

MicroRNA expression arrays and data analysis

Five µg of total RNA from each sample was sent to the OSU Comprehensive Cancer Center MicroArray Shared Resource for miRNA expression analysis using the miRNA microarray chip OSUCCC version 4.0 (Liu et al., 2004; Liu et al., 2008). For each strain of mice, a fourth RNA sample of a second independent RNA isolation was also evaluated. The arrays contain over 300 human and mouse miRNAs spotted in triplicate with probes for both precursor and mature miRNAs. Experiments were performed and analyzed as previously described (Liu et al., 2004; Liu et al., 2008). In brief, a linear model was employed to detect differentially expressed miRNAs. In order to improve the estimates of variability and statistical tests for differential expression, a variance smoothing method with fully moderated t-statistic was employed for this study (Yu et al., 2011). The significance level was adjusted by controlling the mean number of false positives (Gordon et al., 2007).

Quantitative PCR

Quantitative PCR (qPCR) validation of candidate miRNAs. qPCR was performed using Applied Biosystems TaqMan probes to validate expression of the six candidate miRNAs (Foster City, CA). Probes were specific to either mature (miR-1, miR-133a, miR-124a-3, miR-206) or precursor (miR-134, miR-206, miR-9-1) miRNAs concordant with the array results. Reverse transcription for mature miRNAs was performed using Applied Biosystem’s High Capacity cDNA Reverse Transcription reagents according to manufacturer’s protocol in 7.5 µl reactions (Carlsbad, CA). Three independent skin RNA samples were used for each strain from samples not used on the original arrays. qPCR of miR-1 levels from eleven primary mouse cSCCs was conducted as for normal skin. For miRNA quantification, real-time PCR was performed using Applied Biosystem TaqMan Assays per manufacturer’s recommended conditions. All samples were done in triplicate. qPCR runs included no template and no-reverse transcriptase controls. Expression was normalized to expression levels for sno202 RNA for mature miRNAs or to L19 for pri-miRNAs. qPCR was conducted using Applied Biosystems 7900HT instrument. Cycle threshold (CT) values were averaged across triplicates and delta CT values were calculated between each test miRNA and control. The fold expression of SPRET/EiJ miRNA relative to FVB/NJ was calculated. Standard deviations of fold differences of the percentage of reference gene expression were made by comparisons between independent RNA samples of each of the test samples. Student’s t-test was used to calculate p-values for the qPCR. A nonparametric Mann-Whitney U test was performed to test for differences between miR-1 expression in tumors versus SPRET/EiJ and FVB normal skin.

qPCR of predicted mRNA targets. To measure expression of predicted targets, mRNA expression was measured at 48 and 72 h post miR-1 or scrambled precursor miR negative control transfection in A5 cells and at 48, 72 and 96 h post-transfection in B9 cells using SYBR green quantitative real-time PCR. Reverse transcription of mRNA was performed using a Bio-Rad iScript cDNA synthesis kit according to manufacturer’s recommended conditions. Primers were designed using Integrated DNA Technologies. Target gene expression was measured using Bio-Rad SYBR Green Supermix in 10 µl reactions according to manufacturer’s protocol. qPCR of each sample was performed in triplicate and was used to measure Ets1, Met, Bag4, Twf1, Sp1, Taok1, Zfp148 and Trp53 target gene expression. The following primer sets were used. Ets1: Forward: 5’ TGTATGAGTGGAGCAGCACTGTGT 3’, Reverse: 5’ AGGTAGGGTCTCCATTAACCT 3’, Met: Forward: 5’ AACGGGTATTGGGAAGACCCTGAA 3’, Reverse: 5’ ATCCCGTCTAACAGGAAGAAGGCT 3’, Bag4: Forward: 5’ ACTCCACGGAAGTTCCAA ACACCT 3’, Reverse: 5’ TTCCAGGGTTCTGTGAAGCAGGAT 3’, Twf1: Forward 5’ TTCCAGGCTTTGGAGAAGGTGAGT 3’, Reverse: 5’ AGTCTCCTTCGTGGGAATGCTTGT 3’, Sp1: Forward 5’ TGCGGCAAAGTATATGGCAAGACC 3’, Reverse 5’ ACTCCTCATGAAACGCTTAGGGCA 3’, Taok1: Forward 5’ GAACCAGGCCAAGTGAAACTTGCT 3’, Reverse 5’ AAAGGAGGCTTCCTCTCGGCTAAT 3’, Zfp148: Forward 5’ GCACTGCAATGCTGCCTTTAGAAC 3’, Reverse 5’ TTTCGCCGGTATGGATCTTCTCGT 3’, Trp53: Forward 5’ ACAAGAAGTCACAGCACATGACGG 3’, Reverse 5’ TTCCTTCCACCCGGATAAGATGCT 3’, and L19: Forward: 5’ ATTCCCGGGCTCGTTGCCGGAAAA 3’, Reverse: 5’ ATTGGCAGTACCCTTCCTCTTCCCTA 3’. Candidate targets were normalized to L19 expression and fold difference relative to the scrambled control at 48 h post-transfection were calculated. Standard deviations of fold differences of the percentage of reference gene expression were made by comparisons between independent RNA samples. Experiments included no template and no reverse transcriptase controls for each gene. Student’s t-test was used to calculate p-values.

qPCR confirmation of transfections. RNA was isolated from a duplicate well for each mock, scrambled control or miR-1 transfected experiment to confirm increased miR-1 expression in the miR-1 transfected cells. qPCR for miR-1 expression was conducted as described above.

Identification of candidate target genes for differentially expressed miRNAs

Target prediction programs, www.microRNA.org (Beta et al., 2008) and www.targetscan.org were searched for predicted mRNA targets for miR-1 in both mouse and human. Targets for miR-1 were prioritized for further study by the number of predicted binding sites per 3’UTR (>1 site), strength of binding score, and predicted binding of the miRNA in both human and mouse. A literature scan was also performed to identify targets which had previously been validated experimentally.

Cell culture and transient transfections

A5 and B9, mouse cutaneous spindle and cSCC cell lines respectively, and C5N, a non-tumorigenic mouse keratinocyte cell line, were used for these studies (Zoumpourlis et al., 2003). Cell lines were grown in Dulbecco’s medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin in an atmosphere of 5% CO₂. Transient transfections were performed using Invitrogen Lipofectamine 2000 according to manufacturer’s protocol (Life Technologies, Grand Island, NY). Mock and negative control (scrambled pre-miR) transfections were carried out for all experiments when cells were at 40–50% confluency. Scrambled pre-miR (AM17110) negative control and miR-1 (AM17150) precursor were purchased from Ambion (Life Technologies, Grand Island, NY).

Western blot analysis

Protein from whole cell extracts was extracted 24, 48, 72 and 96 h post-transfection via solubilization of A5 and B9 cells using RIPA buffer (50 mM Tris base pH 8, 150 mM NaCl, 1% NP40, 0.10% SDS) containing a protease inhibitor (Roche). For each sample, 30 µg of protein were run on a 10% polyacrylamide gel in a Bio-Rad mini-gel system at 150 V (12 V/cm) for ∼1.5 h. Proteins were transferred to nitrocellulose membrane for 70 min at 90 V (7 V/cm). Following transfer, membranes were blocked in 5% milk in TBST (10X TBS: 1.2% Tris, 8.8% NaCl) and were incubated with primary antibody overnight. Antibodies and dilutions were as follows: Ets1, 1:1000 dilution (antibody a gift from Michael Ostrowski) and α-tubulin, 1:100 dilution (T5168, Sigma-Aldrich, St. Louis, MO). Incubation with secondary antibodies was for 2 h. Secondary antibodies and dilutions were as follows: Anti-mouse HRP-Linked, 1:10000 dilution (W4021, Promega, Madison, WI), Anti-Rabbit HRP-Linked Antibody, 1:1000 dilution, (7074S, Cell Signaling Technology, Danvers, MA). Enhanced chemiluminescent reagent (Thermo Scientific, Rockford, IL) was applied in a 1:1 ratio.

Luciferase assays

Cloning. Cloning of Ets1 3’UTR was performed using In-Fusion Advantage PCR Cloning Kit according to manufacturer’s protocol (Clontech, Mountain View, CA). Full length Ets1 plasmid with 3’UTR provided by Dr. Michael Ostrowski was used as a template for PCR. A 1,278 base pair product was generated containing the predicted three miR-1 binding sites and was cloned into a pGL3 Control Luciferase vector (Promega, Madison, WI). Primers used for cloning were as follows: Ets1 3’ UTR In-Fusion Forward: 5’-tgtaatactagtccgAGAAGAGAGGCAATTGGCTGAGGT-3’, Ets1 3’ UTR In- Fusion Reverse: 5’-gtctgctcgaagcggACTGAGGCAGTATTCCTGATAGAG-3’. Clones were sequence verified. Site-directed mutagenesis was carried out to mutate two base pairs in each of the three miR-1 binding sites in the Ets1 3’UTR using QuikChange Lightning Multi Site-Directed Mutagenesis kit according to the manufacturer’s protocol (Agilent Technologies, Santa Clara, CA). Primers used for site-direct mutagenesis are as follows:

Ets1 3’ UTR SDM 1 Forward: 5’-gttgatggctgactcccactctcccttgaagactctgaat-3’,

Ets1 3’ UTR SDM-1 Reverse: 5’-attcagagtcttcaagggagagtgggagtcagccatcaac-3’,

Ets1 3’ UTR SDM-2 Forward: 5’-ccacaaggaagcaaaggccaaactctccagctatatattttgatctta-3’ ,

Ets1 3’ UTR SDM-2 Reverse: 5’-taagatcaaaatatatagctggagagtttggcctttgcttccttgtgg-3’,

Ets1 3’ UTR SDM-3 Forward: 5’-acctgttgaactcttacgtactctccaaagacgtttcaaggaac-3’,

Ets1 3’ UTR SDM-3 Reverse: 5’-gttccttgaaacgtctttggagagtacgtaagagttcaacaggt-3’.

Clones were sequenced verified.

Transfections and luciferase assays. C5N and A5 cells were plated in triplicate into a 12-well dish at 24 h prior to transfection. At 70% confluency, 0.10 µg Luc-Ets1 3’UTR and Luc-Ets1 3’UTR-mutated were each co-transfected with 10 pmol of a scrambled control miR or miR-1 into each well. 0.10 µg of a PRL-TK vector (TK-driven Renilla luciferase vector, Promega) was also transfected into each well to normalize the firefly luciferase values. All experiments included mock and pGL3-Control empty vector controls. A5 and C5N cell lysates were prepared using M-PER (Pierce Biotechnology, Rockford, IL) and 30 µg of each sample were used for analysis. Firefly and renilla luciferase measurements were acquired 24 h post-transfection using a Veristas Microplate Luminometer.

Cell proliferation assay

At 24 h post-transfection with miR-1, scrambled miR control or mock transfected, A5 and B9 cells were trypsinized and 2000 cells (A5) or 3000 cells (B9) were plated in quadruplicate in a 96-well plate. Proliferation was measured at 24, 48, 72 and 96 h post-transfection using a MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) cell proliferation kit (Roche, Indianapolis, IN). Experiments were carried out according to manufacturer’s protocol. RNA and protein were isolated at each solubilization step. Optimal absorbance of the formazan product was measured at a wavelength of 550 nm and a reference wavelength of 690 nm was used. For each sample, absorbance was normalized to the reference wavelength as well as a blank control, which contained media only.

Cell cycle assay

A5 and B9 cells were transfected with miR-1, scrambled miR control or mock transfected. Cells were trypsinized and counted 48, 72 and 96 h post-transfection. Two million cells were fixed in 70% ethanol. Cell cycle analysis was evaluated by DNA content. Cells were incubated with a propidium iodide solution (0.1% Triton X-100, 0.2 mg/mL DNAse-free RNAse A and 20 µg/mL propidium iodide) for 15 min at 37 °C. Data was acquired on a BD FACS Calibur instrument. ModFit software was used for analysis. Experiments were done in triplicate.

Apoptosis assay

A5 and B9 cells transfected with miR-1 or scrambled miR control were trypsinized and counted at 48, 72 and 96 h post-transfection. The FITC Annexin V Apoptosis Detection Kit I was used to evaluate cell death according to manufacturer’s protocol (BD Biosciences, San Jose, CA). Unstained, FITC Annexin only, and propidium iodide only conditions were used as controls. Data was acquired on a BD FACS Calibur instrument. Cell Quest Pro for was used for data analysis. Experiments were done in triplicate.

Cell motility assay

A5 and B9 cells were plated at a density of 7 × 10⁵ cells/mL in ibidi cell culture inserts according to the manufacturer’s protocol (Martinsried, Germany) at approximately 48, 72 and 96 h post-transfection. Transfections of A5 and B9 cells included mock transfection, scrambled miR control (AM17110) or miR-1 precursor (AM17150) as described above. Eight hours after plating, cell inserts were removed. Images were taken at 5X magnification using AxioVision 4.8 software on an Axiovert 25 microscope at 0 h and every 5–8 h until gaps closed. Images were assessed visually by two independent researchers for differences in rate of gap closure.

miRNA profiling of normal skin of cancer resistant and cancer susceptible mice

We hypothesized that miRNAs differentially expressed between cancer resistant and cancer susceptible mice might have oncogenic- or tumor suppressor-like properties in the skin. To identify miRNAs that were differentially expressed between cancer resistant SPRET/EiJ and cancer susceptible FVB/NJ mice, we performed miRNA expression profiling of total RNA isolated from four normal skin samples per strain. The miRNA arrays contained probes for precursor and mature miRNAs of approximately 300 miRNAs (Table S1; Liu et al., 2004; Liu et al., 2008). Following analysis, six miRNAs (miR-1, miR-133a, miR-124a-3, miR-134, miR-206, and miR-9-1) showed a significant difference in average expression and had a 2.0 fold or greater difference across one or more probe sets. All of these showed higher expression in SPRET/EiJ (Table 1). Due to the evolutionary divergence between M. musculus, for whom the majority of arrays have been designed, and M. spretus, lower miRNA expression observed in SPRET/EiJ mice could indicate a sequence polymorphism. However, we observed the opposite effect in which all miRNAs showing differential expression were higher in SPRET/EiJ. These results decrease the likelihood that the observed profiles were due to polymorphic variants between the strains.

To validate the expression differences of the six miRNAs showing differential expression between SPRET/EiJ and FVB/NJ by microarray, we performed qPCR using Applied Biosystems TaqMan probes on samples from mice not used in the initial array studies. Probes were specific to either mature (miR-1, miR-133a, miR-124a-3, miR-206) or precursor (miR-9, miR-134, miR-206) miRNAs and were chosen for evaluation based on the specific probe sets (mature versus precursor) that showed differences by microarray (Table 1). Five of the six miRNAs evaluated showed similar fold-differences in expression between microarray and qPCR for FVB/NJ and SPRET/EiJ skin (Fig. 1 and data not shown); however, only miR-1, miR-133a and miR-206 showed statistically significant differences in expression by qPCR (p-value < 0.05).

miRNA-1 expression decreased in mouse cSCCs

We hypothesized that miRNAs expressed at higher levels in SPRET/EiJ mice relative to FVB/NJ mice would behave similarly to tumor suppressor genes and exhibit reduced expression in tumors. We chose miR-1 for expression studies in mouse cSCCs to test our hypothesis because it showed a significant difference in expression between FVB/NJ and SPRET/EiJ for six probes on the array and showed a significant difference in expression by qPCR (Table 1). In addition, there was evidence from the literature showing down-regulation of miR-1 in a variety of human cancers and a link between miR-1 down-regulation and cancer phenotypes (Nasser et al., 2008; Datta et al., 2008; Yan et al., 2009; Nohata et al., 2012a; Nohata et al., 2012b). From these data, we expected that miR-1 expression would be decreased in mouse cSCCs. To evaluate miR-1 expression in cSCC, we performed qPCR of miR-1 in eleven cSCCs isolated from DMBA/TPA chemically-treated C57Bl6/FVB mice. The median miR-1 expression was lower in the tumors (3% of control sno-202) compared to median miR-1 expression in FVB/NJ (22% of control sno-202) and SPRET/EiJ (110% of control sno-202). The difference in expression was significant between the tumors and SPRET/EiJ (Fig. 2, p-value = 0.012). Based on the expression data in the primary tumors and the strain specific differences in expression, we chose miR-1 as a strong candidate miRNA to evaluate further for a role in cSCC.

qPCR of miR-1 candidate target genes

miRNAs are thought to impact gene regulation by targeting mRNAs for degradation via complete sequence matches or inhibition of translation when there is imperfect binding between a miRNA and its target sequence (Djuranovic, Nahvi & Green, 2011). We hypothesized that if the difference in miRNA expression between the strains of mice was important in cancer susceptibility there would be differences in target mRNA expression of genes associated with carcinogenesis. To identify putative targets of miR-1, miR-133a, miR-124a-3, miR-134, and miR-206, we searched the literature to identify targets of these miRNAs that had been detected by expression arrays in tumors and had been further validated by other methods. We also identified genes predicted to be targets in microRNA databases www.microRNA.org (Beta et al., 2008) and www.targetscan.org. We prioritized genes as candidates which had two or more predicted binding sites, had stronger binding scores and those genes previously implicated in cancer. Using these strategies, we identified multiple candidate targets for each miRNA (Table S2).

We chose seven putative miR-1 targets, Bag4, Ets1, Met, Sp1, Taok1, Trp53, and Zfp148, and one positive control, Twf1, which previously had been shown to be a miR-1 target by Applied Biosystems, for testing. These candidate target genes were prioritized based on our database searches and those which had been validated in the literature as being targets of miR-1 in studies of primary human tumors or cancer cell lines. By qPCR, two of the candidate targets, Ets1 and Met, and our positive control Twf1 showed significant down-regulation in miR-1 transfected A5 cells compared to scrambled miR precursor transfected cells at both 48 and 72 h post-transfection (Fig. 3 and Fig. S1, Met p-values 0.0001 and 0.002; Ets1 p-value = 0.0001 and 0.0002; Twf1 p-values = 0.0001 and 0.006 respectively). Our positive control, Twf1, showed decreased expression in miR-1 transfected B9 cells at 48, 72 and 96 h (p-values = 0.0004, 0.0002 and 0.03 respectively). In B9 miR-1 transfected cells Met showed significantly decreased expression only at 48 h (p-value = 0.001), and Ets1 showed significantly decreased expression at 48 and 72 h post-transfection compared to scramble miR control transfected cells (p-values 0.008 and 0.005 respectively). Bag4 mRNA expression was significantly reduced in B9 cells transfected with miR-1 at all three time points (Fig. S1; 48 h p-value = 0.0007, 72 h p-value = 0.02 and 96 h p-value = 0.004), but it only showed significant down-regulation of mRNA at 72 h in A5 cells transfected with miR-1 compared to scrambled control cells (Fig. 3D and Fig. S1, p-value = 0.01). The remaining targets did not show statistically significant differences in expression between control and miR-1 A5 transfected cells (data not shown).

Evaluation of Ets1 as a miR-1 target

mRNA expression data is only correlative and does not show a direct effect of miRNAs on predicted target gene expression. As Ets1 had not been confirmed as a target of miR-1 in the mouse, and since several studies support a role of Ets1 in cSCC and cancer phenotypes such as metastasis, apoptosis and proliferation, we decided to further evaluate murine Ets1 as a miR-1 target (Keehn, Smoller & Morgan, 2004; Hahne et al., 2009; Nagarajan et al., 2009). We first assessed whether Ets1 protein levels were decreased by miR-1 expression. By Western, Ets1 was down-regulated in the miR-1 transfected A5 cells at 24, 48 and 72 h post-transfection and in B9 cells at 48 and 96 h post-transfection compared to scramble control miR transfected cells (Figs. 4A and 4B; Fig. S2). As this could be an indirect effect of miR-1 expression, we assessed the effect of miR-1 expression on a luciferase construct containing the 3’UTR of Ets1. miR-1 expression resulted in lower luciferase expression with the wildtype Ets1 3’UTR compared to a scrambled precursor miRNA. Furthermore a construct containing mutated miR-1 binding sites did not show reduced luciferase expression providing additional evidence that miR-1 regulates Ets1 expression via direct binding to the predicted miR-1 binding sites in the Ets1 3’UTR (Fig. 4C). Our data is consistent with a publication showing Ets1 to be a target of miR-1 in a human liver cancer cell line (Wei et al., 2012).

Functional characterization of miR-1 expression

miR-1 expression in cSCC cells decreases cell proliferation

There are several published studies showing in vitro and in vivo effects of miR-1 on tumor suppression (Nasser et al., 2008; Datta et al., 2008; Yan et al., 2009; Nohata et al., 2011; Nohata et al., 2012a; Nohata et al., 2012b). Based on these, we anticipated that expression of miR-1 in cSCC cells would have tumor suppressing abilities (Nohata et al., 2012b). To determine if miR-1 decreased cell proliferation, we transfected A5 and B9 cSCC cells with miR-1 or scrambled precursor miRNA. Cell proliferation was measured by an MTT assay at 24, 48, 72 and 96 h post-transfection. We observed a significant decrease of 1.7-fold in absorbance at 96 h post-transfection in A5 miR-1 transfected cells compared to the scrambled control miR transfected cells suggesting that miR-1 is playing a role in growth inhibition (Fig. 5A, 72 h p-value = 0.01; 96 h p-value = 0.0001). A modest, but significant decrease in proliferation was observed at the 72 h time point in the B9 cSCC cell line, but there was no difference in proliferation at 96 h (Fig. 5B, 72 h p-value = 0.02; 96 h p-value = 0.3).

The effect of miR-1 expression on additional tumor phenotypes

miR-1 has been shown to induce apoptosis in multiple cancer cell lines including maxillary sinus SCC, head and neck SCC, and renal cell carcinoma (Nohata et al., 2011; Kawakami et al., 2012; Nohata et al., 2012a; Nohata et al., 2012b). To determine if the decreased proliferation seen in the cells transfected with miR-1 was the result of increased apoptosis, we performed FACS analysis of AnnexinV and PI stained miR-1 transfected and scramble control miR transfected A5 and B9 cells. No significant differences in apoptosis were observed at 48 or 72 h post-transfection in the A5 cells; however, there was a significant increase in apoptosis at 96 h (p-value = 0.036; Fig. 6A and Fig. S3). The B9 cells showed a statistically significant decrease in apoptosis in the miR-1 transfected cells at 48 h (p-value = 0.04) and a statistically significant increase in apoptosis at 72 h (p-value = 0.003). There were no differences in apoptosis at 96 h (Fig. 6B and Fig. S3). The apoptosis data for the A5 and B9 cell lines are consistent with the proliferation results.

Ectopic expression of miR-1 has also been shown to induce G0/G1 arrest in several different cancer cell lines (Nohata et al., 2012b). We evaluated cell cycle parameters in miR-1 transfected cells by FACS analysis (Fig. 6 and Fig. S4). We observed modest but consistent and statistically significant fewer cells in the S phase for miR-1 transfected A5 cells compared to scrambled control miR transfected A5 cells at 48, 72, and 96 h (p-values = 0.002, 0.0001 and 0.006 respectively), Figs. 6C, 6E and 6G and for miR-1 transfected B9 cells at 48 and 96 h (p-values = 0.001 and 0.007 respectively), Figs. 6D and 6H. There was a trend for a modest increase in cells in G0/G1 for miR-1 transfected A5 cells at all time points, but this was only significant at 48 h post-transfection (p-value = 0.04); Fig. 6C. This was not observed in the B9 cells. There was a trend for more cells in G2-M in the miR-1 transfected cells for both A5 and B9 for some time points (Fig. 6 and Fig. S4), but the absolute differences are quite small and were not significant except for the miR-1 transfected B9 cells at 48 h (p-value = 0.02), Fig. 6F. These results suggest that miR-1 may have an extremely modest effect on cell cycle progression resulting in fewer cells in S phase.

Previous studies have reported a role of miR-1 in migration. In addition, Ets1 has a well-documented role in enhancement of migration (Hahne et al., 2005; Smith et al., 2012). To test whether miR-1 affects migration in cSCC, we performed cell motility assays of miR-1 transfected and scrambled control transfected A5 and B9 cells at approximately 24, 48, 72 and 96 h post-transfection. No consistent differences were observed in cell motility between miR-1 and scrambled control miR transfected cells for any of the time points (Fig. 7 and Fig. S5 and data not shown). Closure of gap appeared to be directly proportional to gap width.

miR-1 expression following transfection was confirmed for all phenotypic experiments by qPCR and is below detectable limits for the mock or scrambled control transfected A5 and B9 cell lines.