The miR-183 family cluster alters zinc homeostasis in benign prostate cells, organoids and prostate cancer xenografts

The miR-183 cluster, which is comprised of paralogous miRs-183, -96 and -182, is overexpressed in many cancers, including prostate adenocarcinoma (PCa). Prior studies showed that overexpression of individual pre-miRs-182, -96 and -183 in prostate cells decreased zinc import, which is a characteristic feature of PCa tumours. Zinc is concentrated in healthy prostate 10-fold higher than any other tissue, and an >80% decrease in zinc is observed in PCa specimens. Here, we studied the effect of overexpression of the entire 4.8 kb miR-183 family cluster, including the intergenic region which contains highly conserved genomic regions, in prostate cells. This resulted in overexpression of mature miR-183 family miRs at levels that mimic cancer-related changes. Overexpression of the miR-183 cluster reduced zinc transporter and intracellular zinc levels in benign prostate cells, PCa xenografts and fresh prostate epithelial organoids. Microarray analysis of miR-183 family cluster overexpression in prostate cells showed an enrichment for cancer-related pathways including adhesion, migration and wound healing. An active secondary transcription start site was identified within the intergenic region of the miR-183 cluster, which may regulate expression of miR-182. Taken together, this study shows that physiologically relevant expression of the miR-183 family regulates zinc levels and carcinogenic pathways in prostate cells.


Results
Physiologically relevant overexpression of the miR-183 cluster in benign prostate epithelial cells. In order to assess the functional effects of the miR-183 family cluster when expressed from its endogenous genomic sequence at physiologically relevant expression levels, the entire 4.8 kb miR-183 cluster genomic region containing miRs-183, -96 and -182 (miR-183 full cluster or 183FC) was cloned into CD511B-1, a lentiviral vector with a GFP tracer (Fig. 1A). Transduction and GFP + selection (Fig. 1B) of benign prostate epithelial RWPE-1 cells with CD511B-183FC (RWPE1-183FC) resulted in 10-30 fold overexpression of the miR-183 cluster members compared to CD511B-CTRL transduced cells (RWPE1-CTRL) population (Fig. 1C). These levels are similar to those observed in prostate cancers compared to benign prostate 18 . Of note, the expression levels were determined by absolute quantitation and expression of the three miRs was not uniform, suggesting differential regulation of the three miRs. Transfection with luciferase constructs containing miR consensus sites in the 3′ UTR validated that the overexpressed miRs were functional (Fig. 1D).

ZIP1 repression and reduced total and accessible zinc in RWPE1-183FC cells. The ZIP1 3′
UTR contains two validated binding sites for the miR-183 family 18 . A 50% and 53% reduction in the levels of ZIP1 mRNA and protein, respectively, was observed in the RWPE1-183FC cells compared to RWPE1-CTRL ( Fig. 2A,B). Total zinc levels were reduced significantly by 15% in RWPE1-183FC cells compared to RWPE1-CTRL by a chromogenic assay (Fig. 2C).
The spatial distribution of accessible zinc was assessed with Zinquin ethyl ester (ZQ), a fluorescent cell-permeable zinc probe that becomes trapped within the cell after cleavage of the ethyl ester moiety by cytosolic esterases 20 . Naïve RWPE-1 cells were used as a control and were mixed 1:1 with the GFP + RWPE1-CTRL or RWPE1-183FC cells. Confocal microscopy showed a perinuclear and cytoplasmic distribution of zinc that was dramatically reduced in the GFP + RWPE1-183FC cells compared to naive RWPE-1 cells (Fig. 2D), whereas ZQ in the GFP + RWPE1-CTRL cells was similar to the naïve cells. Digital image quantification of individual cells showed that ZQ intensity was significantly lower (*p < 0.0001) in the RWPE1-183FC cells compared with the in RWPE1-CTRL and RWPE1-183FC cells. N = 3 experiments were analyzed as percentage of CTRL mean per experiment. Data from all 3 experiments were analyzed together by Mann-Whitney nonparametric test and are shown as boxplots with mean (line), 25-75% (box) and 95% confidence intervals (whiskers) and individual cell outliers. *p < 0.05 by Mann-Whitney nonparametric test for non-normal distribution of data in parts A and E, Student's paired 2-sided t-test for parts B and C.
RWPE1-CTRL GFP + cells (Fig. 2E). These data show that miR-183 cluster overexpressing cells have reduced levels of intracellular zinc. Notably, the zinc phenotype is striking in the ZQ study as ZQ binds to free (labile), loosely-bound zinc, and accessible zinc 20 .
Primary human epithelial prostate organoids with 183FC had diminished zinc by x-ray fluorescence. Primary human prostatic epithelial cells (PrE), derived in our lab, were used as a second in vitro preclinical model to assess zinc regulation by 183FC. Following lentiviral infection, single cell PrE cells were cultured in matrigel for 14 days to form prostate organoids ( Fig. 3 and Supplemental Fig. 1). 183FC organoids were markedly smaller than the GFP controls (Fig. 3A). Total zinc was assessed by X-ray fluorescence (Fig. 3B,C and Supplemental Fig. 1) and was lower in 183FC organoids. Notably, the 183FC organoids lacked zinc in the differentiated cells in the centres of the organoids (Fig. 3C). This reduction in zinc was similar in magnitude to the reduction of zinc in PCa tissue compared to benign patient tissue by the same method (Fig. 3D).
To test the effects of the miR-183 cluster family in vivo, RWPE2-CTRL or RWPE2-183FC cells were grown as subcutaneous xenografts in male NUDE mice. Whereas the RWPE2-CTRL tumours reached a maximum tumour volume of 40 mm 3 then spontaneously regressed to 20 mm 3 , the RWPE2-183FC continued to increase in size until the study end at 13 weeks (Fig. 4G). Moreover, intra-tumoural zinc levels were significantly lower in the RWPE2-183FC grafts compared to the RWPE2-CTRL xenografts (Fig. 4H). These experiments show that the miR-183 cluster family augments tumour growth of RWPE-2 cells and also reduces zinc in vivo. miR-183 cluster regulates cancer-related pathways. Gene expression was profiled in the RWPE1-CTRL and RWPE1-183FC by microarray. Genes significantly lower in the 183FC included those involved in adhesion, angiogenesis and metabolism (Fig. 5A). Gene set enrichment analysis and DAVID Pathway analysis showed enrichment of beta cell, RNA pol 1 promoter opening, metabolism and adhesion gene sets (Supplemental Table 1). Of note, hZIP1 (SLC39A1) was 82% lower in the 183FC compared to CTRL (data not shown), validating the RT-qPCR and immunoblot data in Fig. 2.
The phenotypic effects of the miR-183 family cluster were examined in the RWPE-1 and RWPE-2 -CTRL and -183FC cells as well as PC3 cells with reduced miR-183 family activity. PC3 cells were derived from castration resistant prostate cancer and represent highly malignant PCa cells. We previously reported that PC3 cells have the highest levels of miRs-183, 96 and 182 of the prostate cancer cell lines 18 . Thus, PC3 cells were transduced with a sequence complementary to the miR-183 family to block activity (PC3-183KD). Cell adhesion was increased in RWPE1-183FC cells compared to RWPE1-CTRL, but there was no change in the other cell lines (Fig. 5B). Cell migration was reduced in the PC3-183KD cells (Fig. 5C,D) and these cells also had decreased suppression of the miR-183 family 3′ UTRs (Fig. 5E). Migration in the RWPE-1 and RWPE-2 cells was minimal and not able to be assessed between the cell lines (data not shown).
The intergenic region between miR-96 and -182 contains an active transcription site to regulate expression of miR-182. Previous RT-qPCR data from our lab showed differential expression by RT-qPCR of the three miRs-183, -96 and -182 with miR-182 being the highest in prostate cells 18 , suggesting a differences in the mature miR transcription, processing and/or stability. As well, Figs 1C and 4C show miR-182 expressed at higher levels than the other 2 miRs by absolute quantitation 22 . We further examined the comparative expression of the miR-183 family members in PrE cells (n = 2) and laser-capture micro-dissected benign prostate epithelium (n = 8) by small RNA-sequencing (Fig. 6A,B). The RNA-sequencing data were consistent between the PrE cells and benign epithelium in showing that miR-182 was present at 10-100 fold higher levels than miRs-183 and -96 (Fig. 6A,B).
To test whether the intergenic region (IGR) between miR-96 and -182 had promoter activity that may be driving higher expression of miR-182, the IGR was cloned into pGL4 as the promoter (Fig. 6C). The IGR had significant promoter activity in all prostate cell lines examined (Fig. 6C). An −1500 bp upstream promoter has previously been reported 23 and we cloned it into pGL4. The IGR promoter activity was 10-20% of the upstream promotor ( Fig. 6C), suggesting that both promoters regulate the expression of miR-182. Targeted deletion of the IGR showed a dramatic decrease in promoter activity when the region 1-2 kb from miR-182 was deleted (Fig. 6D). Interestingly, the highest level of sequence conservation in the IGR also lies in this 1-2 kb region, supporting the presence of regulatory elements in this region that further regulate miR-182 separately from miRs-183 and -96.

Discussion
The goal of this study was to investigate whether overexpression of the miR-183 family from its native gene cluster decreased intracellular zinc levels and altered PCa growth. Though the phenomenon of low zinc in PCa has been documented for many decades, it is still unclear whether loss of zinc contributes to prostate carcinogenesis or whether it is a bystander in the development of the tumour. The miR-183 family was previously reported by our group to be highly expressed in prostate cancers and was a regulator of zinc homeostasis 18 . However, a limitation of our prior study, and of most miR studies, was that the miRs were overexpressed by transient transfection pre-miR mimics which results in supra-physiological levels (often >1000 fold) of the mature miRs and bypasses potential regulation during pri-miR transcript processing 18 . The present study avoids the caveats of pre-miRs by cloning the entire 4.8 kb miR-183 cluster into a lentiviral vector for expression in cells to preserve the highly conserved putative regulatory elements within the 4.2 kb intergenic region between miRs-96 and -182. This method for overexpression is a more natural mimic of increased expression of all three of the miRs that is observed in PCa.
Our findings not only validated our previous work on zinc regulation, but also demonstrate that these phenotypes are observed at levels of the miR-183 family cluster that mimic those observed in tumours. The overexpression of the miRs decreased ZIP1 protein and mRNA expression in the miR-183FC transduced RWPE-1 and RWPE-2 compared to their respective controls. Importantly, zinc regulation by 183FC similarly occurred in primary epithelial prostate organoids and in vivo in RWPE-2 xenografts, demonstrating that the role of the miRs is not unique to one cell type or an in vitro phenomenon. The RWPE2-183FC tumours were also larger than controls, which consistent with other reports of overexpression of the miRs individually in PCa cell xenografts [24][25][26][27] . The decrease in zinc observed in the 183FC tumours further supports the link between zinc and PCa carcinogenesis. Since zinc depletion has been such a consistent feature in PCa, it is not clear whether the low zinc is permissive of cancer growth or is simply decreased due to epithelium dedifferentiation and loss of the zinc sequestering phenotype. Franklin and others have provided strong rationale that lower zinc does contribute to carcinogenesis due to reactivation of the Krebs cycle 1, 3, 12 .
Our ZQ data suggest that the miR-183 family members have a pronounced effect on free, loosely-bound or accessible zinc. The reduction in ZQ intensity was ~30% compared to a 15% reduction in total zinc as measured by both the Quantichrome and X-ray fluorescence. It is not possible to directly compare the assays statistically, however, the two total zinc methods (Quantichrome and XRF) had highly similar results, providing some confidence in this difference. Although this difference may seem minimal, this labile pool of free, loosely-bound or accessible zinc that is capable of affecting the cell metabolism via inhibition of aconitase as shown by Franklin and Costello 1-3 . Therefore it may be that changes in expression levels of the miR-183 family are integral for the loss of zinc observed in PCa and subsequently alters the metabolism.
Overexpression of the miR-183 family also regulated cell phenotypes such as migration, adhesion and proliferation. These phenotypes differed between benign and PCa-derived cells. In benign cells, high levels of 183FC did not alter the proliferation of RWPE-1 cells and actually reduced the size of benign PrE organoids, suggesting proliferation changes require further molecular changes. As well, since PrE cell organoids develop from progenitor cells, it is possible that the lower levels of zinc in the 183FC organoids inhibited differentiation. Gene expression analysis of RWPE1-183FC cells suggested regulatory control of several key pathways in carcinogenesis including adhesion and motility. Concordantly overexpression of 183FC enhanced adhesion in RWPE-1. Migration of all RWPE cell lines was minimal (>48 hr scratch closure), however knockdown of the miR-183 family in the more motile PC-3 cells significantly decreased their motility. In the PCa RWPE-2 cells 183FC enhanced cell proliferation in vitro and xenograft tumour volume in vivo.
The miR-183 family are considered "oncomiRs" in many other cancers and the miRs have been shown to regulate pathways dysregulated in cancer such as DNA repair, apoptosis and metabolism [28][29][30][31][32][33][34][35][36][37][38][39][40] . In our study, the genes with the highest magnitude of suppression were adhesion related, and this has not previously been reported. There are several potential explanations to our observation. The first is that the lowered zinc phenotype that accompanies miR-183FC overexpression is prostate cell specific, leading to repression of different target genes. Another possibility is that the coordinated expression of the miRs as a cluster (including the 4.2 kb IGR) provides a different target profile from overexpression of a single miR, and in this way our study is unique from others that have profiled gene regulation by the miRs. As further evidence of a regulatory role for the IGR, we observed promotor activity of the sequence in all prostate cell lines. Given our array findings and the promoter activity of the IGR, the role of the 4.2 kb IGR needs to be studied further as this region possesses highly conserved areas in eukaryotes that may further regulate the expression of the miRs in this cluster.
Our study adds to the understanding of miR function and regulation in prostate physiology and as a potential contributor to PCa. We have much left to learn about the function of miRs in a cell-type specific manner and transcriptional/post-transcriptional mechanisms that govern miR expression itself. Our findings may not be exclusive to PCa and could extend to other cancers that overexpress the miR-183 family. Although the zinc phenotype of PCa is well-described, it does not exclude the possibility that other tumour types have zinc disruptions. It is also important to distinguish that our work investigates the molecular changes leading to changes in cellular zinc and is not meant to advocate for dietary zinc interventions, which have had mixed results on PCa risk [41][42][43] . Cloning and lentiviral production. miR-183 cluster (4.8 kb) was isolated from human male DNA by nested PCR with primers F outer 5′-AAGGTCATCTTGGGCTGATG-3′ and R outer 5′-GAGGGTGAGAAGGGTCACCA-3′ and F nested 5′-GACTGAAGTCGGGGGTAGAGACCGTAGCAG -3 ′ an d R ne ste d 5 ′ -C TG AC T TC AG G G ATG C AG G G A A AC AC AG AG -3 ′ . T h i s P C R pro duct (miR-183FC) was cloned into an intermediate vector and sub-cloned using primers wit h N heI sites: 5′ AT TC G A AT TC GCTAGCT GGCTGTGCACAGGGTGC 3 ′ and 5′ AGATTCTAGAGCTAGCGGGAACGGGCATCGTGGG 3′. miR-183FC was cloned into lentiviral vector CD511B-1 (Systems Biosciences, Mountain View, CA) with NheI enzyme (New England Biolabs, Ipswich, MA) and propogated in Stellar Competent Cells (Clontech, Mountain View, CA). Clones were verified by sequencing. Lentiviral transduction. Ultrahigh titre of lentivirus from CD511B-1 + miR-183FC and CD511B-1 scrambled control was produced by Systems Biosciences (Mountain View, CA) and used to transduce RWPE-1, RWPE-2, and PrE cells by spinfection 45 via UIC Institutional Biosafety Committee approved protocol (#15-020). Briefly, lentivirus (120 MOI) and polybrene (8 μg/ml) (Sigma Aldrich St. Louis, MO) were added to 75,000 cells in a polystyrene tube, incubated 1hr 37 °C, centrifuged at 750 × g 25 °C for 1hr and re-plated. GFP expression was evident at 72hr using EVOS fluorescence microscope (ThermoScientific). PC-3 with knockdown of the miR-183 family were generated using the miRZIP Lentivector-based Anti-miR system (System Biosciences, Palo Alto, CA), with shRNA targeted to miR-183 or control scrambled shRNA, and maintained under selection with 1.5 μg/ml puromycin.

Sorting of GFP + population of cells. The transduced RWPE cells were GFP + cells were sorted on MoFlo
Astrios (Beckman Coulter, Brea, CA) in 1 ml of FACS sorting buffer (1X PBS +0.5% BSA +1 mM EDTA) based on GFP detection by 488 nm laser. Sorted cells were collected into collection buffer (1X complete KSFM +1X Ab/ Am +30% FBS), washed and plated to expand the cultures. miR and mRNA expression. RNA was isolated from cells at 70% confluency using Trizol (ThermoScientific) with the small RNA recovery protocol. RNA quantity and quality was determined with the NanoDrop Spectrophotometer (Thermo Scientific, Waltham, MA) by absorbance at 230, 260 and 280 nm. For miR RT, 150 ng of RNA was converted to cDNA using the Exiqon Universal cDNA kit (Exiqon, Denmark). For the mRNA-RT, 355 ng of RNA was converted to cDNA using High Capacity Reverse Transcription Kit (ThermoScientific). SYBR real-time RT-qPCR was run on the QuantStudio6 (ThermoScientific). C T values were normalised to housekeeping RNAs (as indicated in the figure legends) by the ΔΔC T method 46 .
For absolute quantitation, RNA oligos with 5′ phosphorylation for miR-183, -96 and -182 (IDT) were used as an RT template to generate a standard curve to calculate the number of miRs per ng of total RNA in the reaction. RNU44 was used to normalize for input.
Luciferase-3′ UTR for miR activity. LightSwitch-RenSP vectors from SwitchGear Genomics (Carlsbad, CA, US) for control (3′ UTR with no miR consensus sites) and miRs-96, -182 and 183 (miR-specific sites in 3′ UTR) were used. Cells were plated at 30 K (RWPE) or 50 K (PC-3) cells/well in a 24-well plate, grown overnight, then transfected with 100ng of LightSwitch vector and empty pGL4-luc control vector with DharmaFECT (Pittsburgh, PA, USA). Renilla and luciferase activity were assayed in triplicate 24 h after transfection with Dual-luciferase assay (Promega, Madison, WI, USA). 3′ UTR levels were normalised to the LightSwitch control vector for each cell type.
Insoluble fraction was pelleted at 10000 × g for 10 minutes (4 °C). Thirty µg of protein was separated on a 4-12% Bis-Tris NuPage gel (Thermo Scientific) and transferred to PVDF membrane. ZIP1 chicken polyclonal antibody (gift from Dr. Renty Franklin) was used at 1:10000 and β-tubulin rabbit antibody at 1:1000 4 °C overnight. Secondary anti-chicken or anti-rabbit (Li-cor, Lincoln, NE) was used at 1:1000 for 1 hour at room temperature and was analysed on the Li-cor Odyssey system. Total intracellular zinc. Cells were harvested into zinc-depleted hypotonic buffer (10 mM Tris, 10 mM NaCl, 3 mM MgCl 2 , 1% NP-40, 10 mM PMSF chelated with chelex). For subcutaneous tumours, the frozen tissues were homogenised in 200 µl of chelated hypotonic buffer. Cell and tissue samples were sonicated, centrifuged at 10000xg (4 °C) and the supernatant assayed with the Quantichrome Zinc Assay kit (BioAssay Systems, Hayward, CA). Zinc was normalised to total protein which was determined by Biorad Protein Assay (Bio-Rad, Hercules, CA).
Zinquin fluorescence of free and loosely bound zinc. Naïve RWPE-1 cells were mixed 1:1 with GFP + RWPE1-CTRL or RWPE1-183FC and allowed to attach overnight. Zinquin (20 µM)(Santa Cruz Biotechnology, Dallas, TX) was added and incubated for 20 minutes at 37 °C. Cells were imaged using confocal microscopy (Zeiss, Ontario, CA) and the EVOS FL Cell Imaging System (ThermoFisher, USA). GFP + cells were compared to GFP − naïve cells with inForm Cell Analysis Software (Perkin Elmer, USA) followed by Mann-Whitney U nonparametric test of the values using GraphPad Prism.
Gene expression profiling of 183C target genes. RNA (2 µg) from RWPE1-CTRL and RWPE1-183FC biologic triplicates were labelled and hybridised to GeneChip ® Human Gene 1.0 ST Arrays (Affymetrix, Santa Clara, CA, USA). Raw cel files were deposited into GEO (GSE99134) Analysis. RMA normalization was performed for the raw CEL files using the R Bioconductor "oligo" package 47 . The well annotated genes were kept using the annotation file from Affymetrix (HuGene-2_0-st). GSEA 48 was used to identify the gene sets (pathways) associated with RWPE1-CTRL and RWPE1-183FC using the curated gene sets (C2) from molecular signature database (MSigDB).
To identify the differentially expressed genes (DEGs), genes with median expression levels <the 20th percentile of medians for all genes or with variance <the 20th percentile of variances for genes were removed. The R Bioconductor "limma" package 49 was used to obtain the top differentially expressed genes between RWPE1-CTRL and RWPE1-183FC, using a false-discovery rate (FDR) of less than 0.2 and absolute fold change greater than 1.5 as thresholds. Pathway annotation for the list of differentially expressed genes was done with the "Functional Annotation" module in the Database for Annotation, Visualization and Integrated Discovery (DAVID) 50 .
Zinc measurement by X-Ray fluorescence of primary prostate organoids and prostate tissues. Benign PrE cells were derived from radical prostatectomy patients via an Institutional Review Board (IRB)-approved human subjects Biorepository protocol with informed consent as previously described 44,51,52 . All experiments with human cells and tissues were performed in accordance with the human subjects' guidelines and regulations in a de-identified manner. Benign PrE cells were transduced with CTRL or 183FC by spinfection and plated in growth factor reduced matrigel ® (Corning, USA) 1:1 with PrEGM. After 14 days, intact organoids were removed with dispase (1 U/mL) and fixed in 4% paraformaldehyde. The organoids were visualised on the EVOS FL Cell Imaging System and GFP + organoids were collected by pipet. Organoids were placed in histogel and paraffin-embedded. Sections (10 µm) were placed onto 2 mm × 2 mm silicon nitride membrane windows (Silson Ltd, Northampton, England).
Prostate tissues were acquired from radical prostatectomy patients via the same IRB-approved biorepository protocol with consent. Fresh tissues were fixed in 4% paraformaldehyde, paraffin-embedded and sectioned onto the windows. X-ray fluorescence data were collected at the X-ray Science Division beamlines at the Advanced Photon Source Beamline 2I-DE, Argonne National Laboratory (Lemont, IL). Data visualization and quantitation of regions of interest (ROI) was performed with MapsSy (Stefon Vogt, Argonne National Lab). Three ROI were drawn for each organoid, the background was subtracted and the mean μg/cm 2 for zinc was normalised to the mean sulfur for each slide. Data from all organoids (six CTRL and four 183FC) from two different patients were included in student's t-test analysis.

RWPE-2 subcutaneous grafts.
The animal experimental protocol (#12-076) was approved by the Institutional Animal Care and Use Committee (IACUC) at UIC, also known as the Animal Care Committee (ACC). All experiments were performed in accordance with US federal guidelines for animal care. RWPE2-CTRL or RWPE2-183FC (7.5 × 10 5 cells) were mixed 1:1 with matrigel ® and injected into the flank of 17 week old athymic foxn/nu nude mice (Harlan, Indianapolis, IN). A silastic testosterone implant (1 cm) was placed subcutaneously in the mice as previously described by Hu et al. 53 . Tumour size was measured weekly using calipers. Tumour volumes were calculated using the following formula: ½ × (Length × Width × Width). Mice were euthanised and the tumours were snap frozen in liquid nitrogen for zinc measurements.
Scientific RepoRts | 7: 7704 | DOI:10.1038/s41598-017-07979-y Small RNA sequencing of PrE cells and benign prostate epithelium. RNA was collected from PrE cells as described above. For the tissues, RNA was isolated with RNAqueous ® -Micro kit (Ambion, USA) from benign prostate epithelium which was collected by laser-capture micro-dissection from eight de-identified frozen prostatectomy specimens (UI Biorepository) [54][55][56] . RNA quality was assessed by the Bioanalyzer (Agilent) and libraries were constructed with 500 ng of RNA using the TruSeq Small RNA Sample Prep Kit (Illumina, USA). The libraries were quantitated by qPCR and sequenced 8/lane for 51 cycles on a HiSeq. 2500 using TruSeq Rapid SBS sequencing chemistry v2. Fastq files were generated and demultiplexed with the bcl2fastq v1.8.4 Conversion Software (Illumina). The number of reads per patient ranged from 14 to 21 × 10 6 . Analysis. Adapters were trimmed, the reads reduced and mature miRs were mapped to hg19. Number of reads for each miR is shown as a sum of all isomers and normalised to raw reads.
Luciferase Assay for promoter activity. The intergenic region between miRs-96 and -182 was isolated by PCR from gDNA with primers F-GAGACTCGAGGGAGCCCAGCAATCTGAG and R-TCTCAAGCTTCTCGGTCTGTGCTGAGGAA. The region was cloned in pGL4 (Promega, Madison, WI, USA) and verified by sequencing as pGL4-IGR. Deletion mutants of with various lengths of promoter upstream of miR-182 (3.5 kb: 3,524 bp; 3 kb: 3,014 bp; 2 kb: 2,021 bp; 1 kb: 1,059 bp and 5 kb: 613 bp) were cloned into pGL4 (Creative Biogene, Shirley, NY). Cells were plated in a 24-well plate in triplicate and transfected using DharmaFect with pGL4 empty vector or pGL4-IGR and prL-null (Promega) as a control. Luciferase and renilla values were measured using the Dual-luciferase ® Reporter Assay System (Promega) on the Glomax 20/20 (Promega).
Luciferase relative light units (RLU) were normalised to renilla for transfection and cell number control.
Migration assay. Cells were grown to confluence in a 24-well plate (3 replicates), then scratched with a Falcon 100 µl pipette tip down the centre of each well. Images of the same location in each scratch were taken using the EVOS FL Auto (Thermo Scientific) at times 0 hr, 12 hr and 24 hr post-scratch. Scratch area was measured using inForm 2.2 software (Perkin Elmer).
Adhesion assay. Cells in media containing NucBlue Live Reagent (ThermoScientific) were plated at 5,000 cell/well in a 96-well plate for 15 m, 30 m, 45 m and 1hr total (3 replicates each) before the plate was inverted and tapped to remove media from the wells. Adherent cells were imaged using the EVOS FL Auto (ThermoScientific). As a control for total cells plated, separate wells were imaged at 3hr without media removal. Cell counts were generated using inForm 2.2 software (Perkin Elmer).
Statistics. Statistics were performed using GraphPad Prism 5.04. Most data were normally distributed and were analyzed by Student's 2-sided t-test; paired for experimental replicates and unpaired for xenograft study. Relative expression by qPCR and Zinquin fluorescence were not normally distributed and were analyzed by Mann-Whitney non-parametric test. p < 0.05 was considered significant. Experiment-specific details are in the figure legends.