Supplementary Figure S1 SNAI1 induces mesenchymal genes prior to repression of epithelial genes, concomitant with miR-424 upregulation. (A) Phase contrast images of HMLE Empty Vector and HMLE SNAI1-inducible cells after 4 and 12 days of SNAI1 induction. Bar-100 μm. Epithelial and mesenchymal markers, by qPCR, after (B) 4 and (C) 12 days of SNAI1 induction, two-tailed unpaired t-test. (D) Mature and (E) primary miR-424 levels, by qPCR, after 4 days of SNAI1 induction, two-way ANOVA, Bonferroni multiple comparison post-test. SEM shown, representative images/graphs of n≥3. #p<0.1, *p<0.05, **p<0.01, ***p<0.001.
Supplementary Figure S2 MiR-424 drives an intermediate EMT that does not repress epithelial programming. (A) Phase contrast images of SUM149PT cells stably expressing miR-424 and (B) miR-200c levels, by qPCR. MCF12A cells stably expressing miR-424 examined for (C) N-cadherin protein, (D) morphology and (E) miR-424 levels and (F) EMT markers, by qPCR. HMLE cells stably expressing miR-424 examined for (G) morphology, (H) miR-424 levels and (I) EMT markers, by qPCR, and (J) N-cadherin protein. SEM shown, n≥3. *p<0.05, **p<0.01, ***p<0.001, two-tailed unpaired t-test. Bar-100 μm.
Supplementary Figure S3 MiR-424 increases classical EMT phenotypes while decreasing TIC phenotypes. Stable miR-424 expressing MCF12A cells displaying: (A) distance migrated over 8 hours; (B) cell cycle analysis, by BrdU/PI staining; (C) adhesion to fibronectin-coated wells; and (D) ALDH1 activity quantification. SEM shown, n≥3. #p<0.1, *p<0.05, ***p<0.001, two-tailed unpaired t-test.
Supplementary Figure S4 Inducible miR-424 recapitulates constitutive miR-424 EMT and TIC phenotypes reversibly. SUM149PT clones with i-EV or i-miR-424 constructs were assayed for a variety of EMT parameters: (A) 3’ UTR luciferase reporter with perfect or mutant miR-424 binding sites after a doxycycline induction (induction shown in Figure 2H) followed by doxycycline removal; (B) cell cycle analysis, by PI staining (G1 shown); (C) gap closure at eight hours; (D) flow cytometry for the TIC marker ALDH1. Doxycycline and reversed conditions were treated with 0.5 μg/mL doxycycline for 7-10 days, while reversed conditions were cultured for an additional 7-10 days in the absence of doxycycline. SEM, shown. *p<0.05, **p<0.01, ***p<0.001. Two-way ANOVA, Bonferroni multiple comparison post-test for all.
Supplementary Figure S5Global mature miR levels are not downregulated in metastatic tissue. (A) Levels of mature miR-25, -125b and -195 in SUM149PT pre-injection cells expressing EV control or miR-424. Levels of mature (B) miR-25, (C) miR-125b and (D) miR-195 in metastatic tissue formed from EV or miR-424 expressing SUM149PT cells, by qPCR. SD shown. &No remaining RNA to assay.
Supplementary Figure S6 MiR-424 is post-transcriptionally regulated in vitro. Primary and mature miR-424 levels, by qPCR, in (A) HMLE cells constitutively expressing SNAI1 and in (B) HMLER cells constitutively expressing TWIST1 or SNAI1, after passaging the cell lines for longer than 20 days. SEM shown, n≥3. *p<0.05, **p<0.01. (A) Two-tailed unpaired t-test and (B) one-way ANOVA with Bonferroni post-test.
Supplementary Figure S7 MiR-424 regulates many genes associated with EMT and cancer stemness. Effects of MCF12A cells stably expressing constitutive miR-424 on (A) known downregulated genes in an EMT signature (1) and on (B) known upregulated genes in a breast cancer stem cell signature (2), as demonstrated using RNA-seq with GSEA. TargetScan-predicted miR-424 binding sites (blue, asterisk) are statistically enriched in the downregulated genes, as assessed using a hypergeometric test with multiple correction permutations (A: p<10-5, B: p=0.002). Levels of TGFBR3, by qPCR, in (C) HMLE and (D) MCF12A cells stably expressing miR-424. SEM shown. **p<0.01, two-tailed unpaired t-test. (E) Flow cytometry for TGFBR3 after transfection of a rescue construct into SUM149PT cells expressing miR-424.
Supplementary Figure S8 Working model. MiR-424 can drive an intermediate EMT in primary breast tumors by downregulating TGFBR3, leading to increased mesenchymal characteristics concomitant with unaffected epithelial content. Experimental motility is also increased that may help facilitate tumor cell dissemination, although TIC phenotypes are repressed. Downregulation of TGFBR3 by miR-424 leads to increased ERK signaling that is required for inhibition of TIC phenotypes, which is reversible in vitro, and we hypothesize is reversed in disseminated tumor cells by decreased miR-424 to facilitate colonization.
Supplementary Figure S9 Expanded materials and methods.
Cell culture
Stable, constitutive miR-424 overexpression was achieved through lentiviral infection. Lentiviral plasmids (pCDH-CMV-miR-424-EF1-Puro) were purchased (System Biosciences) and packaged according to manufacturer’s recommendations. Infected cells were selected with puromycin (SUM149PT – 1 μg/mL; HMLE – 0.5 μg/mL; MCF12A – 2 μg/mL). TGFBR3 knockdown was achieved with Mission TRC2 shRNAs (Sigma; TRCN0000359000, TRCN0000359081) obtained through the UCCC Functional Genomics Facility. Virus production and infection were completed as described above. Expression of TGFBR3 was achieved by transiently transfecting pcDNA3.1 HA-TGFBR3, a generous gift from Dr. Gerard Blobe (Duke University), with FuGene 6 (Promega) according to manufacturer’s instructions.
Inducible miR-424 expression was achieved by cloning 131 bp upstream and 11 bp downstream of the pre-miR-424 stem-loop sequence (Accession #MI0001446, mirbase.org) out of HMLE gDNA into the XhoI and MluI sites, respectively, of pTRIPZ. EV and miR-424 pTRIPZ were packaged into virus particles as described above and infected into SUM149PT cells expressing firefly luciferase. Cells were selected with puromycin and single-cell sorted for turboRFP negative cells to grow out clones without a miR-424 expression leak.
GSEA and human data analyses
GSEA (3) was performed as previously described (4) with our RNA-seq data and using GeneSigDB as the gene sets (5). Contingency table analysis used GSE22220, comprised of 207 primary breast cancers (6), and GSE19536, comprised of 91 primary breast cancers (7). Patient GSEA studies also used GSE22220. GSE41970, comprised of 64 triple-negative breast cancers (8), was used for matched tissue analysis. A hypergeometric test was performed to determine the significance of predicted miR-424 target genes in GSEA results. In addition, a permutation test by randomly resampling 10,000 times was performed to determine the p-value of predicted miR-424 target genes. P-value < 0.05 was considered statistically significant.
TCGA clinical information was obtained from the TCGA website (9). Processed and normalized data and subtype information for 291 breast cancer samples were downloaded from cBio (10). TCGA samples were identified as high or low miR-424 expression (above or below median miR-424 expression). MiR-424 expression levels were compared in different breast cancer subtypes using an ANOVA test and corrected by Tukey HSD, using R statistical program (R Foundation for Statistical Computing), with p-value < 0.05 considered statistically significant.
Quantitative PCR
Total RNA was purified using miRNeasy Mini Kit (Qiagen), and cDNA was made using miScript II RT (Qiagen). SsoFast-EvaGreen and SsoFast-Probes Supermix (Bio-Rad) were used for dye and probe-based qPCR, respectively. QPCR samples were run on a CFX-96 Real-Time PCR Detection System (Bio-Rad) and analyzed with CFX Manager Software (v2.0; Bio-Rad) using the ddCt method. GAPDH and RNU6-2 (U6) were used as reference genes for mRNAs and miRs, respectively. All kits and reagents were used according to manufacturers’ recommendations.
QPCR Primer sequences
qPCR Target / Primer Sequence/SourceFibronectin / F-GAGTGTGTGTGTCTTGGTAATGG
R-CCACGTTTCTCCGACCAC
E-cadherin / F-TTACAGTCAAAAGGCCTC
R-AGCGTGACTTTGGTGGAA
Plakoglobin / F-CAGATCATGCGTAACTACAG
R-CACACGGATAGCACCTT
α-catenin / F-TTGGCTGCATCTGTTGAA
R-GGCTCTCCTTCGCAATT
TGFBR3 / F-TGGAGTCTCCTCTGAATGGCTG
R-CCATTATCACCTGACTCCAGATC
GAPDH, N-cadherin, Vimentin / Mani et al. Cell. 2008.
miR-424, -200c, -25, -125b, -195, U6 / Qiagen
pri-miR-424 / F-CCTTCATTGACTCCGAGGGGA
R-CGGCAGACCCCACCTTCTA
Microscopy
Phase contrast photos were taken on a CKX41 microscope (Olympus) using DP2-BSW software (v2.2; Olympus). Immunofluorescence photos were taken on an Axiovert-200M (Zeiss). Images were deconvoluted by nearest neighbor analysis and Z-stacks were collapsed using Slidebook (v5.0; Intelligent Imaging Innovations).
Mouse models
For tumor-initiation experiments, cells were diluted 1:1 in complete media:Matrigel (BD Biosciences) for a total volume of 100 μL per sample. 6-week old female NOD/SCID mice were anesthetized with 2.5% isoflurane, and were injected underneath the nipple of the #4 mammary fat pad with cell numbers as outlined in the text, as previously described (4). Tumor formation was monitored by weekly palpation.
For experimental metastasis assays, 6-week old female nude mice were anesthetized with 2.5% isoflurane, and 5x105 luciferase-tagged cells, in 100 μL PBS, were injected into the left heart ventricle as described (11). Immediately after intracardiac injection of cells, mice were imaged by in vivo imaging system (IVIS) for firefly luciferase to observe systemic tumor cell disbursement, ensuring proper injection into the left ventricle. Mice failing to meet this criterion were removed from the study. Metastatic burden was then monitored over time by IVIS as previously described (11), and metastasis incidence was determined by persistent IVIS signal in the same location over multiple time points. To isolate metastases, ex vivo IVIS imaging was used to confirm the site of metastatic outgrowth and to remove as much normal tissue as possible, before mouse tumor tissue was placed in Qiazol (Qiagen) and homogenized before being processed as described in the Quantitative PCR section.
3’ UTR reporter
Approximately the first 1000 bp of the human TGFBR3 3’ UTR (NM_003243.4) that contained the predicted miR-424 binding site (TargetScan) were cloned into the psiCheck2 luciferase reporter plasmid (Promega) using XhoI and NotI (Primers: F-AAGCCCTCGAGAGCTCAGCTCAGCTACT; R-TATATGCGGCCGCTGAGTGGAAACACTCACC). The miR-424 seed binding site in the psiCheck2 TGFBR3 3’ UTR reporter was mutated from TGCTGCT to ACGACGA (Primers: F-GGGGGTTTTCAATGTGAAACACATGCCAGTTTTTAA AAACGACGATTGTCCAGGTGAGAACATCC; R-GGATGTTCTCACCTGGACAA TCGTCGTTTTTAAAAACTGGCATGTGTTTCACATTGAAAACCCCC). The perfect miR-424 binding site was constructed by annealing oligos exactly complementary to mature miR-424 with oligos comprised of the mature miR-424 sequence. For mutant miR-424 binding sites, similar oligos were used where the seed sequence of CAGCAGCA was mutated to GTCGTCGT.
Expanded Methods References
1.Taube JH, Herschkowitz JI, Komurov K, Zhou AY, Gupta S, Yang J, et al. Core epithelial-to-mesenchymal transition interactome gene-expression signature is associated with claudin-low and metaplastic breast cancer subtypes. Proc Natl Acad Sci U S A 2010;107:15449-54.
2.Creighton CJ, Li X, Landis M, Dixon JM, Neumeister VM, Sjolund A, et al. Residual breast cancers after conventional therapy display mesenchymal as well as tumor-initiating features. Proc Natl Acad Sci U S A 2009;106:13820-5.
3.Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005;102:15545-50.
4.Smith AL, Iwanaga R, Drasin DJ, Micalizzi DS, Vartuli RL, Tan AC, et al. The miR-106b-25 cluster targets Smad7, activates TGF-beta signaling, and induces EMT and tumor initiating cell characteristics downstream of Six1 in human breast cancer. Oncogene 2012;31:5162-71.
5.Culhane AC, Schroder MS, Sultana R, Picard SC, Martinelli EN, Kelly C, et al. GeneSigDB: a manually curated database and resource for analysis of gene expression signatures. Nucleic Acids Res 2012;40:D1060-6.
6.Buffa FM, Camps C, Winchester L, Snell CE, Gee HE, Sheldon H, et al. microRNA-associated progression pathways and potential therapeutic targets identified by integrated mRNA and microRNA expression profiling in breast cancer. Cancer Res 2011;71:5635-45.
7.Enerly E, Steinfeld I, Kleivi K, Leivonen SK, Aure MR, Russnes HG, et al. miRNA-mRNA integrated analysis reveals roles for miRNAs in primary breast tumors. PLoS One 2011;6:e16915.
8.Cascione L, Gasparini P, Lovat F, Carasi S, Pulvirenti A, Ferro A, et al. Integrated microRNA and mRNA signatures associated with survival in triple negative breast cancer. PLoS One 2013;8:e55910.
9.Cancer Genome AtlasNetwork. Comprehensive molecular portraits of human breast tumours. Nature 2012;490:61-70.
10.Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer discovery 2012;2:401-4.
11.Micalizzi DS, Christensen KL, Jedlicka P, Coletta RD, Baron AE, Harrell JC, et al. The Six1 homeoprotein induces human mammary carcinoma cells to undergo epithelial-mesenchymal transition and metastasis in mice through increasing TGF-beta signaling. J Clin Invest 2009;119:2678-90.
1