Chromosome 19 microRNA cluster enhances cell reprogramming by inhibiting epithelial-to-mesenchymal transition

During implantation, cytotrophoblasts undergo epithelial-to-mesenchymal transition (EMT) as they differentiate into invasive extravillous trophoblasts (EVTs). The primate-specific microRNA cluster on chromosome 19 (C19MC) is exclusively expressed in the placenta, embryonic stem cells and certain cancers however, its role in EMT gene regulation is unknown. In situ hybridization for miR-517a/c, a C19MC cistron microRNA, in first trimester human placentas displayed strong expression in villous trophoblasts and a gradual decrease from proximal to distal cell columns as cytotrophoblasts differentiate into invasive EVTs. To investigate the role of C19MC in the regulation of EMT genes, we employed the CRISPR/dCas9 Synergistic Activation Mediator (SAM) system, which induced robust transcriptional activation of the entire C19MC cistron and resulted in suppression of EMT associated genes. Exposure of human iPSCs to hypoxia or differentiation of iPSCs into either cytotrophoblast-stem-like cells or EVT-like cells under hypoxia reduced C19MC expression and increased EMT genes. Furthermore, transcriptional activation of the C19MC cistron induced the expression of OCT4 and FGF4 and accelerated cellular reprogramming. This study establishes the CRISPR/dCas9 SAM as a powerful tool that enables activation of the entire C19MC cistron and uncovers its novel role in suppressing EMT genes critical for maintaining the epithelial cytotrophoblasts stem cell phenotype.


C19MC is expressed in human villous trophoblasts. Previously, laser capture microdissection and
RT-qPCR demonstrated higher C19MC expression in villi-containing regions of human first trimester and term placentas than in EVT regions 13 . However, C19MC expression in situ in the proximal cell columns (proliferative trophoblasts) and in the distal cell columns (invasive trophoblasts) of the anchoring villi at the maternal decidua had not been studied. Thus, we performed in situ hybridizations (ISH) with probes for miR-517a/c, a C19MC miRNA, or scramble control on first trimester and term placental sections. To distinguish trophoblasts (fetal) from decidual cells (maternal), consecutive serial sections were subjected to immunohistochemical staining for cytokeratin and vimentin, respectively.
We observed the strongest miR-517a/c expression in both CTs and STs in the villi, as well as in proximal CTs in the cell columns of the anchoring villi. This miR-517a/c expression gradually decreased in distal CTs with a further reduction in the interstitial EVTs (cytokeratin positive) in the maternal decidua and was undetectable in maternal decidual (vimentin positive) cells. No signal was detected in both ISH and immunohistochemical staining when scramble control or secondary antibody only were used, respectively (Fig. 1a). Term placental sections also displayed strong expression of miR-517a/c in villous trophoblasts (VTs, comprised of CTs and STs), whereas the scrambled control probe showed no staining (Fig. 1b). These results suggest that C19MC is required to maintaining the non-invasive epithelial phenotype of villous trophoblasts. Therefore, we sought to investigated whether C19MC regulates EMT.
Transcriptional activation of C19MC. Investigation of the regulatory role of C19MC on EMT genes, requires an improved technique to efficiently overexpress the entire C19MC cistron in vitro. Hence, we employed the dCas9-based transcription activation system that combines a single-guide RNA (sgRNA) containing two MS2 RNA aptamers, a catalytically inactive Cas9 variant fused to the VP64 gene activator and MS2-p65-HSF1 activation helper proteins, designated as synergistic activation mediator (SAM) system 21 . This system facilitates robust and specific transcriptional activation by using a single targeted guide RNA (sgRNA) while overcoming the constraints of epigenetic regulation of C19MC. We designed two different sgRNAs targeted to the region immediately upstream of the first C19MC miRNAs, the 759-sgRNA binds at ~171 bp upstream of the first two miRNAs of the C19MC, whereas 620-sgRNA binds at ~585 bp upstream of the first miRNA of the C19MC ( Supplementary  Fig. s1).
The expression of 4 randomly selected C19MC miRNAs was validated by RT-qPCR, which showed significant up-regulation of all 4 C19MC miRNAs (300-to 31,000-fold, p ≤ 0.05) in HEK293 cells, while the expression of miR-21, a non-C19MC miRNA, was unchanged further confirming the specificity of the CRISPR/dCas9 SAM system (Fig. 2c). Furthermore, transfection of two human tumor cell lines, MCF7 and PC3 cells with either 759-sgRNA/SAM or 620-sgRNA/SAM also induced C19MC expression ( Supplementary Fig. s2). Lastly, we tested the reproducibility of this method in an immortalized invasive trophoblast cell line, HTR8/SVneo, that does not express C19MC. Since 759-sgRNA/SAM displayed higher C19MC induction efficiency in HEK293 cells, we used 759-sgRNA/SAM and observed 16-to 244-fold increased expression of 4 selected C19MC miRNAs compared to GFP transfected control, despite the low transfection efficiency (Fig. 2d). These results demonstrate that the CRISPR/dCas9 SAM system can be used to selectively and consistently activate the entire C19MC cistron, which spans over 100 kb.
To further investigate the effect of C19MC overexpression on the EMT pathway, the EMT RT 2 profiler PCR Array (Qiagen) was used to quantify changes in the expression of 84 EMT-associated genes and found 2-to 4.75-fold decrease in the expression of CDH2, SERPINE1, TWIST1, SNAI2, CALD1, GSC, ITGAV, MMP3, TCF4 and WNT5a (p ≤ 0.05) in HEK293 cells transfected with 620-sgRNA/SAM compared to control, further confirming the hallmark gene set analysis ( Supplementary Fig. s3). HEK293 cells transfected with 759-sgRNA/SAM or 620-sgRNA/SAM showed significant decrease in CDH2, SERPINE1, TWIST1 and SNAI2 expression compared to control (Fig. 3h). Importantly, HTR8/SVneo cells transfected with 759-sgRNA/SAM also showed decrease in the expression of EMT gene. However, due to their low transfection efficiency (Fig. 2d), the reduction in EMT gene expression did not reach statistical significance (Fig. 3i).
To determine if the decrease in the expression of one of the target mRNAs, specifically CDH2, resulted from direct targeting by C19MC miRNAs, we constructed a luciferase reporter containing the 3′ UTR of the CDH2 gene. Accordingly, HEK293 cells transfected with 759-sgRNA/SAM and the CDH2 3′ UTR luciferase construct displayed a 35% decrease (p < 0.05) in luciferase activity (Fig. 3j). Taken together, these results indicate that overexpression of C19MC induces transcription of major pluripotent factors OCT4 and FGF4 and inhibits EMT by directly repressing expression of genes crucial for EMT.

Hypoxia inhibits C19MC and increases expression of EMT markers.
During the first trimester of human pregnancy, oxygen tension plays a key role in regulating trophoblast cell proliferation and differentiation 23,24 . To evaluate the in vitro effect of hypoxia and avoid the spontaneous differentiation that occurs during culturing of primary human trophoblast cells 25 , we used iPSCs, which share many stem-like characteristics with CTs and express similarly high levels of C19MC. sRNAseq analysis of iPSCs incubated in hypoxia (1% O 2 ) for 24 hours displayed ~ 2-fold decrease in the expression of C19MC (mir-498 cistron), while the hypoxia-inducible miR-210 26 was increased 13.9-fold (FDR ≤ 0.20) compared to normoxia (21% O 2 ) ( Table 1). Specifically, 12 miR-NAs of the C19MC were decreased by 2.1-to 292.6-fold (Supplementary Table S3). Moreover, RT-qPCR analysis confirmed the reduction of five miRNAs of the C19MC by hypoxia (Fig. 4a). www.nature.com/scientificreports www.nature.com/scientificreports/ To determine whether hypoxia regulates genes involved in EMT, we performed RNAseq analysis on iPSCs exposed to hypoxia (1% O 2 ) or normoxia (21% O 2 ) for 24 hours. Hallmark gene set analysis of up-regulated genes revealed that hypoxia induced the expression of genes involved in EMT and P53 pathways (FDR ≤ 0.2), of which SERPINE1 and TWIST1 exhibited 43.7-and 10-fold increased expression, respectively (FDR ≤ 0.20) (Supplementary Table s4). Increased expression of the EMT genes, TWIST, SNAI2 and SERPINE1 was also   www.nature.com/scientificreports www.nature.com/scientificreports/ confirmed by RT-qPCR (Fig. 4b). Furthermore, RT-qPCR analysis of iPSCs exposed to hypoxia displayed down-regulation of the stem cell associated genes LIN28A, LIN28B, SOX2, NANOG, and OCT4 (Fig. 4c).
Hypoxia is known to preferentially promote the initial differentiation of CTs into EVTs 27 . To gain insight into the effect of hypoxia on C19MC and EMT genes during the CT-to-EVT differentiation process, we used a two-step method in vitro that utilizes a low concentration of BMP4 to differentiate iPSCs into CTs 27 (Fig. 4d). After the first differentiation step, iPSC-derived CTs displayed a more flattened and elongated CT phenotype and expressed high levels of the CT associated genes P63 and CDX2 compared to undifferentiated iPSCs (Fig. 4e,f). Moreover, iPSC-derived CTs showed a reduction in OCT4 expression compared to undifferentiated iPSC (Fig. 4f). In the second step, iPSC-derived CTs were cultured in either normoxia (20% oxygen) or hypoxia (1% oxygen) to further differentiate them into STs-and EVT-like cells, respectively 27 . Consistent with previous reports 25, 28 , the expression levels of PSG4 (an ST-associated transcript) was significantly reduced, whereas the expression of HLAG (an EVT marker) was significantly increased when the cells were incubated under hypoxia (Fig. 4g). Importantly, iPSC-derived CTs, or iPSC-derived CTs cultured in normoxia or hypoxia showed a significant increase in the EMT genes CDH2 and SERPINE1 and loss of miR-518c and miR-519d (C19MC miRNAs) compared to undifferentiated iPSCs (Fig. 4h,i). These findings indicate that EMT induced either by hypoxia or by BMP4 used for iPSC-to-CT differentiation is associated with a significant decrease in expression of C19MC.
C19MC enhances cellular reprogramming of somatic cells. Since C19MC miRNAs are highly expressed in embryonic stem cells and its overexpression increased the expression of OCT4 and FGF4, we tested whether the robust activation of the C19MC by CRISPR/dCas9 SAM system induces cell reprogramming. Accordingly, normal human dermal fibroblast cells (NHDFs) were transfected with 759-sgRNA/SAM alone, or with Y4, which consists of three episomal plasmids that contain the Yamanaka reprogramming factors OCT3/4, SOX2, KLF4, and LIN28 in combination with L-MYC and p53 shRNA as previously described 29 , or with a combination of 759-sgRNA/SAM and Y4 (759-sgRNA/SAM +Y4) (Fig. 5a).
NHDFs transfected with 759-sgRNA/SAM alone exhibited a gradual increase in the expression of miR-515-5p (a representative miRNA of the C19MC cistron) that peaked at ~24-fold on day 4 and returned to baseline on day 7 (Supplement Fig. s4). NHDFs transfected with 759-sgRNA/SAM alone did not produce any reprogrammed  (Fig. 5b). Moreover, the colonies were observed earlier in NHDFs transfected with 759-sgRNA/SAM +Y4 at 7 to 10 days post transfection compared to 14 to 20 days post transfection in NHDFs transfected with Y4 alone (Fig. 5c). The undifferentiated stage of the colonies was confirmed by alkaline-phosphatase staining at day 24 post transfection (Fig. 5d). These results indicate that the combination of 759-sgRNA/SAM +Y4 increases both the speed and efficiency of cell reprogramming.
To test the differentiation potential of iPSCs generated by using 759-sgRNA/SAM +Y4 in vitro, we used a floating cultivation method for embryoid body formation. iPSCs generated using 759-sgRNA/SAM +Y4 formed spheroid shaped structures after 8 days in suspension culture (Fig. 5e). Embryoid bodies were transferred to gelatin-coated plates and cultured for an additional 7 days. Attached cells displayed various cell morphologies that included contracting cardiac-like cells and expressed higher levels of the germ layer markers ACTA2 and GATA4 (84-and 37-fold respectively, p ≤ 0.05) compared to undifferentiated control iPSCs (Fig. 5f).

Discussion
Trophoblast differentiation and subsequent invasion into the decidua is critical for implantation and placentation during early human pregnancy and maintenance of pregnancy thereafter. Villous trophoblasts display epithelial characteristics including apicobasal polarity, lateral cell junctions and contact with basement membrane 30 . Upon differentiation to EVTs, trophoblasts undergo EMT, resulting in loss of cell-cell junctions and apical basal polarity and acquisition of a migratory and invasive capacity 30 . The current study demonstrates that C19MC is primarily expressed in the non-invasive villous trophoblasts and is lost as trophoblasts differentiate into EVTs. Under hypoxia, the expression of C19MC is reduced, whereas EMT and cell differentiation are induced. Utilizing the CRISPR/dCas9 SAM system we were able to transcriptionally activate the entire C19MC and demonstrate its role in inhibition of EMT and enhancing the epithelial stem cell state. Moreover, co-expression of C19MC with Yamanaka factors in NHDFs significantly increased the efficiency of somatic cell reprogramming into iPSCs.
The current study establishes in situ, the localization of C19MC cistron in first trimester human placental tissues that is highly expressed in villous trophoblasts comprising of CTs, STs and proliferative proximal trophoblastic cell columns in the anchoring villi. However, as trophoblasts differentiate into EVTs in the distal trophoblastic cell columns and in the interstitial trophoblasts in the decidua C19MC expression appears to diminish. These www.nature.com/scientificreports www.nature.com/scientificreports/ results agree with a previous report that used laser-capture microdissection of the villi-containing regions and EVT regions of paraffin-embedded first trimester placental specimens followed by RT-qPCR 13 .
To transcriptionally activate the entire C19MC cistron, which spans over 100 kb, we employed the CRISPR/ dCas9 SAM system 21 . Comparisons of sRNAseq analysis of three independent experiments using either 759-sgRNA/SAM or 620-sgRNA/SAM in HEK293 cells, revealed that a single guide RNA is sufficient to up-regulate the expression of the entire 100 kb C19MC locus. We observed that 759-sgRNA/SAM, which binds at two locations (~171 bp upstream of the first two C19MC miRNAs), resulted in ~2-fold higher expression of miRNA of the C19MC cistron and higher number of differentially expressed genes compared to the 620-sgRNA/ SAM, which binds to one location (~585 bp upstream of the first C19MC miRNA). These findings are in agreement with previous studies showing that the efficiency of gene editing increases when multiple gRNAs that target the same gene are used 31,32 . The transcriptional activation that we achieved was specific to C19MC since the expression of the neighboring miR-371-373 cluster, which is located 20 kb downstream of the C19MC, was not increased. These results are consistent with a previous report showing that C19MC miRNAs are a product of a single transcript that is rapidly spliced and processed by the Drosha-DGCR8 complex 33 . In the current study, only two up-regulated miRNAs, miR-139 and miR-449, which are not members of the C19MC, overlapped in the 3 independent experiments. Interestingly, these two miRNAs are known to inhibit EMT and may be regulated by the C19MC 34,35 . A few other miRNAs that are known to stimulate cell reprogramming and do not belong to the C19MC were also up-regulated in HEK293 cells transfected with either 759-sgRNA/SAM or 620-sgRNA/SAM. Conversely, eight miRNAs that are known to promote cell differentiation were down-regulated in 620-sgRNA/ SAM and in one of the 759-sgRNA/SAM transfected cells. Additional studies are required to investigate whether these non-C19MC miRNAs are directly or indirectly regulated by the C19MC.
In agreement with the physiological expression of C19MC miRNAs in villous trophoblasts and its gradual loss in EVTs, we found that transcriptional activation of the C19MC led to a significant decrease in the expression of numerous genes involved in Hedgehod signaling (HH) and EMT, several of which are predicted targets of C19MC. These findings are in agreement with a previous study, which showed that BAC mediated C19MC overexpression in HTR8/SVneo cells, decreases their migration 13 . Moreover, several studies have demonstrated a link between HH signaling and EMT 36 . In human trophoblasts, activation of HH signaling induced key EMT regulators including SNAI1 and TWIST1 37 . Therefore, the current results place C19MC upstream of HH and EMT signaling.
The early placental environment is highly hypoxic because intervillous circulation is not established until after the first trimester 23,24 . EVT migration and invasion into the maternal decidua and subsequent remodeling of the maternal spiral arteries are critical steps for proper perfusion of the placenta, which is required for a successful pregnancy 30 . Changes in oxygen tension within the intervillous space are stringently controlled and play a role in trophoblast differentiation, migration and invasion that are essential for vascular remodeling during pregnancy 38 . Previously, primary human trophoblast (PHT) cells were used to test the effect of hypoxia (<1% oxygen) on C19MC expression 39 . However, PHT cells undergo spontaneous differentiation and lose C19MC expression when cultured for 72 hours 39 . To avoid the spontaneous differentiation of PHTs, in this study, we used undifferentiated iPSCs and found that 24 hours exposure to hypoxia (1% oxygen) significantly decreased the expression of numerous C19MC miRNAs and several reprogramming factors including OCT4 and LIN28B, whereas the expression of EMT related genes was increased. The repression of LIN28B expression in iPSCs by hypoxia agrees with our previous study that showed that hypoxia decreased the expression of LIN28B in choriocarcinoma BeWo and JEG3 cells, and in placentas from preeclampsia-affected pregnancies 40 . In addition, repression of OCT4 induces loss of pluripotency and differentiation of ESC cells derived from the inner cell mass toward the trophectoderm lineage 41 . These findings are also consistent with previous studies reporting that hypoxia is a major inducer of EMT [42][43][44] . Although the mechanisms by which hypoxia reduces the expression of C19MC miRNA are not yet known, potential mechanisms include de novo methylation or regulation of C19MC miRNA biogenesis and activity 45,46 . The current study also found significant reduction in the expression of C19MC miRNAs and increased expression of the EMT genes when iPSCs were initially differentiated to bipotential stem-like CTs and subsequently differentiated into STs and EVTs. Although the first step of iPSC differentiation to iPSC-derived CTs increased CT markers, C19MC miRNAs were unexpectedly reduced, perhaps due to the use of BMP4 that induce EMT 41,[47][48][49] . These results indicate that trophoblasts generated using this two-step protocol are atypical since they do not fully recapitulate elevated C19MC expression found in normal trophoblasts. Although our results indicate a strong association between C19MC and EMT, C19MC knockdown experiments in stem cells are required to prove that C19MC regulates EMT.
Our results also reveal that transient activation of C19MC increased the expression of FGF4 and OCT4, significantly increased the number of iPSC colonies when co-transfected with Y4. These data are consistent with a previous study that used a lentiviral vector encoding a C19MC miRNA (mir-524 precursor) 19 . Moreover, some C19MC miRNAs share seed sequences with the miR-302/-372 family, which also enhances reprogramming by promoting MET 50 . Although the mechanisms by which C19MC induces the expression of OCT4 and FGF4 are not yet clear, they may reflect either C19MC mediated inhibition of transcriptional suppressors or binding to the promoter regions and induction of gene expression 51,52 .
The results presented here establish the CRISPR/dCas9 SAM technique as a powerful tool to investigate the role of C19MC cistron in human placental physiology. Employing this robust technique enabled us to uncover the crucial role of C19MC in regulating EMT genes in villous trophoblasts and maintaining their stem-like epithelial cell phenotype. The hypoxic condition during early placentation reduce C19MC expression and releases the inhibition of EMT genes leading to the acquisition of migratory and invasive characteristics of EVTs. Therefore, maintaining optimal expression levels of C19MC is likely critical for EVT differentiation and invasion. Dysregulation of C19MC may result in impaired invasion associated with either the shallow placentation of preeclampsia or the exuberant invasion of placenta accreta.

Materials and Methods
Study approvals. Term placental sections were obtained from pregnancies following vaginal delivery after obtaining a written informed consent and approval by the institutional review board of the University of South Florida (Protocol 00015578). Placental sections from the first trimester (7-and 8-week gestation, n = 2) and early human pregnancies (20-weeks week gestation, n = 2) were obtained from a previously banked deidentified paraffin tissues, under approval by Yale University Human Investigation Committee and by the institutional review board of the University of South Florida. Animal study procedures were approved by the Institutional Animal Care and Use Committee (IACUC, IS00004309) at the Morsani College of Medicine, University of South Florida. All methods were performed in accordance with the relevant guidelines and regulations.
In situ hybridization. In situ hybridization for hsa-miR-517a/c, a C19MC miRNA, was performed in paraffin embedded sections of early pregnancy placentas. After deparaffinization in xylene and rehydration by a series of graded alcohol washes, in situ hybridization was performed using 40 nm 5′,3′ digoxigenin-labeled locked nucleic acid probe for hsa-miR-517a/c (Exiqon, 611715-360) or scrambled (negative) control (Exiqon, 90005). Hybridization and post-hybridization graded SSC washes were performed at 55 °C. The sections were then blocked, and the probes were detected using alkaline phosphatase conjugated sheep anti-digoxigenin Fab fragments (Roche, 11093274910). The signal was developed using NBT/BCIP (Roche, 11697471001) as a substrate that produces dark-blue indigo precipitating dye followed by nuclear counterstaining with Nuclear Fast Red (Vector laboratories, H-3403). The sections were dried and covered with mounting medium for later image analysis.
www.nature.com/scientificreports www.nature.com/scientificreports/ In vitro differentiation of iPSCs. Human iPSCs were harvested by treating with collagenase IV for 30 minutes at 37 °C. Cell clumps were transferred to ultra-low attachment plates (Fisher Scientific, 07-200-601) and maintained in DMEM supplemented with 20% FBS. Media was replaced with fresh medium every other day. After 7 days in suspension culture, embryoid bodies were transferred to gelatin coated 8-well chamber slides and cultured for an additional 10 days. Expression of the different germ layer markers were quantified by RT-qPCR.
Teratoma formation and immunohistochemistry. iPSCs were mixed with Matrigel and injected into each flank of NOD-SCID mice. Tumors were harvested at 4 weeks post-injection, fixed in 4% paraformaldehyde and embedded in paraffin. To differentiate implanted human cells from mouse cells, tumor sections were immunostained with a human anti-mitochondria specific antibody (Abcam cat# ab92824) at 1:1000 dilution followed by secondary antibody (biotinylated horse-anti-mouse, Vector Lab BA-2000, 3.75ug/ml, RRID:AB_2313581) according to the manufacturer's protocol. Nuclei were subsequently counterstained with hematoxylin.
Alkaline phosphatase staining. Alkaline phosphatase histochemical staining was performed using an Alkaline phosphatase detection kit (Millipore, SCR004) as described in the manufacturer's instructions. Briefly, cells were fixed with 4% paraformaldehyde for 10 minutes and washed once with rinse buffer (20 mM Tris-HCl, pH 7.4 and 0.05% Tween 20). Staining solution was added to the wells and plates incubated in the dark for 25 min. Bright field images were then obtained using a light microscope.

RNA isolation and quantitative PCR.
Total RNA was isolated using the RNeasy Mini Kit (Qiagen) and stored at −80 °C in RNAse-free water. For qRT-PCR analysis, 1 µg total RNA was reverse transcribed using random hexamer or oligodT primers and M-MuLV reverse transcriptase (New England Biolabs) according to the manufacturer's specifications. For miRNA cDNA, 0.5 ug of RNA was reverse transcribed using the TaqMan miRNA Reverse Transcription Kit (ThermoFisher Scientific, 4366596) as previously described 11,40 .
Global miRNA-and mRNA-sequencing. Two micrograms of total RNA were converted into a small RNA (sRNA) cDNA library according to published protocol 54 . Briefly, the RNA input for each sample was ligated to a 3′ adaptor barcoded sequence, pooled, size selected, and gel purified, followed by 5′ adapter ligation and then subjected to size selection and gel purification. The cDNA library preparation was completed by second strand synthesis using SuperScript III, alkaline RNA hydrolysis, and PCR amplification for 10 cycles. mRNA libraries were prepared by utilizing the Illumina TruSeq Stranded mRNA LT protocol using 500 ng total RNA and NEB's Protoscript II reverse transcriptase for the first-strand cDNA synthesis according to the manufacturer's protocol. Individual RNAseq libraries were quality controlled on an Agilent TapeStation with a High Sensitivity D1000 ScreenTape. Indexed samples were quantified using the Qubit dsDNA HS assay and were pooled at equimolar concentration (10 nM). The libraries were sequenced on an Illumina NextSeq. 500 sequencer 75-bp paired-end in mid-output mode in the Genomics Core Facility of The Rockefeller University.
Bioinformatics analysis. The miRNA read annotation was performed as previously described 54,55 .
Multimapping reads were assigned to each mapping transcript divided by the number of mapping locations to create a count matrix (i.e. fractional counting). The mRNAseq data were aligned to the human genome build 38 using the STAR aligner 56 (version 2.0.4j) allowing for two mismatches. Expression values (count matrices) were generated using featureCounts together with gene definitions from Ensembl release 82 (GTF file).
Luciferase reporter plasmid, transfection and luciferase assay. The CDH2 3′ UTR was amplified by PCR using the following primer pairs that contain the NOT1 restriction site (F: 5′-ATGCGCGG