Expression and Function of C1orf132 Long- Noncoding RNA, in Breast Cancer Cell Lines and Tissues

Afsaneh Malekzadeh Shafaroudi Department of Biology, Faculty of Science, Ferdowsi University of Mashhad, Mashhad Ali Shari -Zarchi Department of Computer Engineering, Sharif University of Technology, Tehran Saeid Rahmani Department of Computer Engineering, Sharif University of Technology, Tehran Nahid Na si Surgical Department, School of Medicine, Iran University of Medical Sciences, Tehran Seyed Javad Mowla Department of Molecular Genetics, Faculty of Biological Sciences, Tarbiat Modares University, Tehran Andrea Lauria Department of Life Sciences and Systems Biology, University of Turin, Turin, Italy Salvatore Oliviero Department of Life Sciences and Systems Biology, University of Turin, Turin, Italy Maryam Moghadam Matin (  matin@um.ac.ir ) Department of Biology, Faculty of Science, Ferdowsi University of Mashhad, Mashhad


Introduction
Breast cancer (BC) is the leading cause of cancer related death in women worldwide, with an estimated 2,100,000 new cases and 627,000 deaths in 2018 1,2 . In Iran, the age distribution of BC is nearly 10 years lower than their counterparts in many countries 3 . Despite the increased incidence rate, BC is highly curable, if diagnosed and appropriately treated at early stages. Clinically, classi cation of BC is based on histopathological ndings of the patients' tissue samples obtained through surgery or biopsy. This includes the assessment of the receptor status of tumor cells namely estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2) 4,5 . One important subtype of BC is the triple negative BC (TNBC), in which neither of the aforementioned receptors are expressed. TNBC tumors make up 10-30% of all breast cancers and are reported to be the most aggressive form of BC with high histological grade, increased mitotic count, central necrosis and margins of invasion. It is also associated with younger age of incidence and poorer prognosis, in comparison to non-TNBC patterns 6 , as well as an association with shorter duration of breastfeeding, higher age at rst pregnancy and lower parity 7 .
Ductal carcinoma in situ (DCIS) is a pre-invasive form of BC, which 20-50% of them may progress towards invasive ductal carcinoma (IDC) and spread to other tissues during progression of the disease.
However, there is no way to determine which lesions would be stable without treatment and which ones will become invasive and require chemotherapy 8 .
Long non-coding RNAs (lncRNAs) are important but poorly conserved transcribed RNA molecules with at least 200 nt in length which are involved in complex biological as well as pathological processes. Their mechanism of action in gene expression regulation can be summarized in four different ways: I) serving as a signal molecule, II) acting as a molecular decoy and/or sponge, III) guiding some changes in cis or trans gene expression, and IV) acting as scaffolds 9,10 . Dysregulation of several lncRNAs has already been reported in breast cancer. They exert either oncogenic or tumor-suppressive functions with signi cant impacts on various biological and pathological processes, especially in the progression of malignant tumors 11 .
C1orf132 is an intergenic lncRNA located on 1q32.2 between CD34 and CD46 protein coding genes. C1orf132 (also known as MIR29B2CHG) is the host-gene for miR-29b2 and miR-29c with a molecular size more than 20 kb. Its promoter is more likely to be hypermethylated in basal-like breast tumors, comparing to the low grade ones 12 . Bhardwaj and colleagues used next generation RNA sequencing technique on a breast cancer progression cell line model and discovered that the tumor suppressor miR-29c is highly down-regulated during the course of breast cancer cells progression toward TNBC. Moreover, their data revealed that higher miR-29c expression, is associated with a longer survival rate and lower distant metastasis 13 .
Here, we explored the expression pattern of C1orf132 in different types of breast cancer cell lines and tissues. We also discovered a distal promoter area with an enhancing function on miR-29b2/c expression, and an inhibitory effect on cell cycle and cell migration.

Results
C1orf132 locus and its transcripts.
C1orf132, also known as the host gene for miR-29b2 and miR-29c, is located between two protein coding genes, CD34 and CD46 (Fig. 1). There are several transcripts with various lengths and number of exons of MIR29B2CHG located between the two aforementioned coding genes. No ortholog gene for C1orf132 has been reported in mouse, mostly because lncRNAs are not conserved during evolution. For this reason, we searched for potential lncRNAs overlapping with miR-29b2 and miR-29c in mouse genome. Interestingly, we discovered a lncRNA with conserved chromosomal location and transcriptional direction, in reference to its human counterpart (Fig. 1).
The human MiR precursor has 5 or 6 exons with its promoter located (~20 kb) upstream of the miR29b2c ( Fig. 2; indicated with blue color). In addition, we noticed several spliced variants of the gene, some started at apparently different start sites, located upstream of the known promoter for miR-29 host gene variant. The latter nding suggested that the lncRNA could bear some more roles, rather than just being served as a precursor to generate miR-29. Using bioinformatics tools such as Fantom5 supported the existence of at least two distal potential start sites (p1 and p2) for the transcription of long forms of C1orf132, with the putative p2 promoter containing a CpG island ( Fig. 2A-C). ENCODE data exhibited the expression and chromatin modi cations (H3K27ac and H3K4me3) of C1orf132 locus in MCF7 cells, while the active marks were mostly concentrated on two start sites. One for the promoter of miR-29 Page 4/21 precursor and the other one for the potential promoter of the longer transcript ( Fig. 2D-G). According to the ENCODE ChIP-seq data, the latter promoter could interact with more than 80 different transcription factors ( Fig. 2H-J).
C1orf132 was signi cantly down-regulated in TNBC tissue samples.
Expression analysis of different breast cancer subtypes in TCGA revealed a downregulation of the longer transcripts of C1orf132 in TNBC patients. According to this data, the expression level of C1orf132 transcripts originating from p2 promoter were signi cantly decreased in ER/PR negative and TNBC patients, compared to the ER/PR positive as well as non-TNBC ones (Fig. 3A). To con rm this experimentally, we performed qRT-PCR and measured the level of C1orf132 long isoform in tumor samples along with their adjacent non-tumor tissues. The real-time PCR results on 52 paired tissue samples con rmed the same pattern of signi cant (p<0.05) down-regulation of C1orf132 in TNBC vs. non-TNBC samples (Fig. 3B). However, we found no signi cant correlation between this isoform and other subtypes of breast cancer (data not shown).
C1orf132 transcribed from the p2 promoter localizes in the nucleus.
To con rm the potential activity of distal promoters, two reporter constructs were generated, namely pEGFP-1-p1 and pEGFP-1-p2 (Fig. 4A). The constructs were then used for transient transfection of MCF7 cells, along with pEGFP-C1 vector as a positive control (Fig. 4B). The level of transfection e ciency was assessed by the percentage of uorescent cells. The presence of the GFP signal revealed the activity of the putative p2 promoter, in contrast to the predicted p1 promoter, which failed to show any activity ( Fig.   4C, 4D(. Motivated with this nding, we decided to do functional analysis on the p2 promoter by deletion of this region, and to investigate its potential effects on various cellular behaviors as well as transcriptome via RNA-seq technique. To investigate the subcellular localization of C1orf132 in MCF12A cells, we ampli ed the RNA extracts fractioned by Ambion cell fractionation buffer. qRT-PCR data on nuclear and cytoplasmic fractions demonstrated that C1orf132 was prevalently present within the nucleus (p<0.05). U1 and B2M were used as controls for nuclear and cytoplasmic compartments, respectively (Fig. 4E).
Considering the activity of p2 region, we decided to target it with three pairs of guide RNAs (Fig. 5A) after transfecting gRNAs to MCF12A cells, the best sharp PCR product of 438 bps (caused by a genomic deletion of about 2000 bps) belonged to guides number 1&3 (Fig. 5B). Genomic real-time PCR on MCF12A and edited colonies con rmed that the colonies were knocked down and that the colony number 5 had the lowest expression of edited promoter (Fig. 5C).
Flow cytometry analysis revealed a G2/M arrest during the cell cycle progression of the edited compared to unedited MCF12A cells (p<0.05 Fig. 6A and 6B). Data obtained from three different biological repeats demonstrated that the putative p2 promoter knock-down could delay the cell cycle progression and signi cantly increased the number of cells in G2/M phase.
Promoter deletion enhanced the migration ability of MCF12A cells.
As seen in Fig. 6, the initial scratch width was monitored in wound healing experiments in edited (top) vs. unedited MCF12A cells (bottom). The results obtained at time points 0 and 24 h revealed that the migration ability is signi cantly improved in edited cells, compared to the unedited cells (p<0.05; Fig. 6C and 6D).
Gene expression pro ling after C1orf132 knock down The gene expression pro les obtained by RNA-seq in two biological replicates of edited and wild type cells demonstrated 1565 genes with differential expression (adjusted p<0.001), with 551 upregulated and 1014 downregulated genes ( Supplementary Fig. 1A). Moreover, two dimensional PCA graph of the data for the two replicates of C1orf132 edited and unedited MCF12A cells showed a nicely separated replicates of edited cells from wild type cells ( Supplementary Fig. 1B). Accordingly, the same data emerged from the heatmap of differentially expressed genes between edited and unedited MCF12A cells ( Supplementary Fig. 1C).
As lncRNAs could have a general enhancer role on neighboring genes, we rst looked at any potential altered expression of upstream and downstream genes of the edited region. According to the RNA-seq analysis, the two neighboring genes, CD34 and CD46, were not affected in edited cells, however, the expression of mir29b2c precursor declined in edited cells compared to the wild type cells (Fig. 6E). According to GO results, the differential expression analysis of the RNA-seq data revealed several up-and down-regulated genes involved in various biological processes, including the positive regulation of vascular endothelial cell proliferation and regulation of endothelial cell chemotaxis to broblast growth factor biological processes. FGF2 and FGFR1 genes which are responsible for the above-mentioned processes, were upregulated more than 4-fold in edited cells (Fig. 6F). FGF2 is a tumor cell survival factor that helps the cells to escape apoptosis. Moreover, in edited cells the expression fold change for CDH2 (N-Cadherin), as a mesenchymal marker, was upregulated more than twelve folds, compared to the unedited cells. In contrast, the expression of CDH1 (E-Cadherin), as an epithelial marker, was downregulated in edited cells (near zero) compared to the wild type MCF12A cells. Here, ZEB1 as an EMT transcription factor showed about 15-fold upregulation in edited cells (Fig. 6F). ESR1, the gene coding for estrogen recepror alpha (ERα), is under-expressed along with gap junction β-2 (GJB2/CX26) and WNT4 in edited MCF12A cells to near zero. Discussion lncRNAs are the largest class of transcripts in human cells, and also the least understood ones. Among various mechanisms attributed to them, being the host of small RNAs, including miRNAs, is the bestknown function of these regulatory molecules. However, it is not clear if this type of lncRNAs have additional functions other than being simply the host genes for small RNAs. C1orf132 or MIR29B2CHG is the host gene of two miRNAs, miR-29b2 and miR-29c. Here, we investigated if this lncRNA could have other functions in human breast cancer cells.
There are few reports on C1orf132 gene, and its cellular role. It has been reported as one of the ve candidate genes which their CpG islands methylation was used to predict individual's age in forensic medicine 14,15 . Indeed, C1orf132 gene provided the highest accuracy in predicting the age 16 .
In contrast to coding genes, long non-coding genes have low conservation during evolution, which makes it di cult to nd their orthologs in other species. Here, we used a simple logic to nd C1orf132 ortholog gene in mouse by looking at genomic location of mouse miR-29b2 and miR-29c. As we expected, there was a lncRNA gene in mouse hosting both miRNAs. Surprisingly, C1orf132 and its mouse counterpart have an exact chromosomal position, locating between CD34 and CD46 genes, and are transcribed in the same direction, considering their neighboring genes. Logically, there is no need for positional genomic conservation of ortholog miRNA host genes. Indeed, this evolutionary conservation suggests that C1orf132 might have other functions, in addition to being a host gene for miR-29b2c.
Our in silico studies revealed two potential promoters, in addition to the previously known miRNAs precursor promoter 12 , as well as two groups of long and short transcripts for C1orf132. The reason for generating short and long transcripts of the gene is not known, however, we had a similar pattern of short and long transcripts for PSORS1C3 in a previous study 17 . There is a possibility that this kind of gene expression regulation by alternative splicing in a tissue-and cell-speci c manner is a more general feature of lncRNAs and their possible dysregulation could contribute to pathogenesis of many cancers 18 .
To explore potential functions of the long transcripts of C1orf132, we suppressed their expression by excising the distal promoter by means of CRISPR/Cas9 method. It has been already claimed that lncRNAs could have an enhancer effect on their neighboring genes 19 . However, we observed no changes in the expression level of CD34 and CD46 in edited cells, both in our RNA-seq and real-time PCR data; ruling out an enhancer role of these transcripts on their neighboring genes. Interestingly, there existed a signi cant downregulation of miR-29c-3p in edited cells, suggesting an enhancer role of distal promoter on proximal promoter of the gene. The later nding makes it very di cult to interpret the RNA-seq results of the edited cells. Indeed, it is impossible to discriminate the direct effects of C1orf132 from an indirect effect mediated by altered expression of miR-29b2 and miR-29c.
In a bioinformatics analysis of 20 published studies, Yan and colleagues reported that the low expression of miR-29a/b/c is associated with poor prognosis of malignant neoplasms, and could be used as a key biomarker to predict cancers progression and recurrence 20 . Furthermore, these miRNAs exhibit both tumor suppressive and oncogenic roles in different cancers. This might be due to the differences in sample types, location, size, or the period of follow-up study 20 . For example, miR-29c was reported to be upregulated in non-small cell lung cancer (NSCLC) tissues, whereas it was downregulated in lung cancer patients subjected to radiotherapy 21,22 .
We also observed an enhanced migration ability in edited cells, suggesting an invasion suppressor role for C1orf132. This is in agreement with Zhang et al. report in which a signi cant increase of invading cells emerged in miR-29 suppressed HepG2 cells 23 . Jiang et al. reported that miR-29c was remarkably decreased in pancreatic cancer cells and that it had association with shorter overall survival and tumor recurrence in pancreatic cancer patients 24 . It has been well documented that miR-29c is lost as early as the preneoplastic stage of TNBC tumorigenesis, where its downregulation could cause a worse overall survival via regulating several target genes 13 .
EMT is characterized by a loss of function of the E-cadherin adhesion protein (CDH1) in epithelial tissues; and breast cancer is served as a good example of this process. In addition to their role in normal cells, classical cadherins (E-and N-cadherins) have a distinguished role in transforming malignant cells as well as tumor progression, in a process termed 'cadherin switching' 25,26 . This process is regulated by EMT transcription factors, including ZEB family. ZEB1 and ZEB2 are EMT transcription factors, which synergistically increase tumor invasion and cell migration 27 . ZEB1 helps epigenetic silencing of CDH1, by bringing several enzymes to the E-cadherin promoter for epigenetic induction or by inhibiting the expression of stemness-repressing miRNAs to cause dynamic transition of non-cancerous stem cells into cancer stem cells (CSCs) and vise versa 28,29 . In this study, we demonstrated that a cadherin switching evidently took place through a downregulation of CDH1 and an upregulation of CDH2 by almost 12-fold, as well as by almost 14-fold overexpression of ZEB1.
Interaction of N-cadherin (CDH2) and FGF receptor (FGFR) leads to their stabilization on the cell surface.
In this way, N-cadherin increases tumor cell interactions with endothelial and mesenchymal cells 25 . On the other hand, broblast growth factor 2 (FGF2) which can induce proliferation of neoplastic cell and tumor hormone-independent growth 30 , is a tumor cell survival factor, which inhibits cell apoptosis through an autocrine secretory loop. In addition, FGF2 can be an activator of angiogenesis during tumor mass growth and metastasis 31 . In this study, we demonstrated that FGF2 and FGFR1 were differentially over expressed in edited MCF12 cells by almost four folds. By these results, the pathway analysis of our data by Enrichr was nicely validated for up-regulated genes.
On the contrary, we evaluated the expression of ESR1, GJB2 and WNT4 genes using qRT-PCR and found out that these genes are expressed at very low (near to zero) levels, in edited cells. GJB2 and ESR1 with a role in "mammary gland development pathway-pregnancy and lactation" according to WikiPathways 2019 Human were among the downregulated differentially expressed genes in edited vs. wild type MCF12A. Estrogen receptor α (encoded by ESR1), which is expressed in normal breast epithelium has an essential growth and differentiation role and also is associated with the growth and survival of breast epithelial cancer cells 32 as well as a part at progression initiation of estrogen-dependent breast cancers 33 . As mentioned earlier, a signi cant progressive loss of miR-29c, as a result of breast cancer cell line progression model to TNBC, was expected 13 . Therefore, a diminution of ESR1 expression is in line with the miR-29c deduction which we have shown in our edited cells. We also observed a signi cant reduction of GJB2 in our edited cells. GJB2 encodes a member of the gap junction protein family. GJB2 protein is often found in the epithelial cells of the mammary duct (luminal) and can form a gap connection channel in the form of homodimers or heterodimers 34 . This protein in breast cancer was shown to be a tumor-suppressor gene and its very low expression in breast tumor tissues might be associated with hypermethylation 35 .
Our results revealed a down regulation of WNT4 in our C1orf132 knocked-down MCF12A cells. Wnt-4 is a member of the WNT family, which encode important signaling proteins and play a vital role in some developmental and oncological processes. Several reports have shown the association of Wnt-4 to epithelial cells, like its expression in normal murine keratinocytes 36 and also in the epidermal cells throughout skin development 37 as well as its expression necessity for normal mammary gland development 38 . Seitoh and colleagues reported that in keratinocyte cell lines, a poor differentiation and more malignant phenotype is related to the loss of Wnt-4 gene expression. In another study, a similar association between Wnt4 and mesenchymal to epithelial transition during kidney development has been nicely shown 39 . In contrast, Vouyovitch and colleagues reported that depletion of WNT4 in MCF12 cells inhibited cellular proliferation, and upregulated WNT4 in non-malignant breast cells, stimulated growth, inhibited apoptosis and increased cell migration and EMT 40 .
As the sole reference on C1orf132 lncRNA and cancer, Peng et al. detected two lncRNAs (C1orf132 and TMPO-AS) as a prognostic signature for lung adenocarcinoma patients, which was validated in two independent datasets, and also by the GO enrichment analysis. By pathway analysis, they suggested that down regulation of C1orf132 was associated with a poor prognosis in lung cancer patients, probably by deregulating the "cell cycle and cell adhesion molecules" pathways in cancer cells 41  Clinical tissue samples.
In this study, a total number of 52 pathological tissue samples with a diagnosis of primary breast cancer were collected. None of the patients had been treated with preoperative radiotherapy, chemotherapy, or other relevant modalities. Breast cancer tissue samples along with their matched adjacent apparently normal tissues were collected, immediately preserved in liquid nitrogen, and then stored at -80ºC until analysis. All patients signed informed consents and agreed to the use of their surgical specimens for research.

Bioinformatics analysis.
We used UCSC genome browser to scan the genomic area around C1orf132 (located at chr1: 207,978,592-208,052,441 (hg19)) for potential promoter activity. Histone modi cations by ChIP-seq from ENCODE (H3K27ac and H3K4me3) and DNA-seq were used to detect the active area near transcription start site (TSS). Using ENCODE ChIP-seq data for different transcription factors (TFs), a list of TFs which bind to C1orf132 potential promoters was extracted. Fantom5 data were used to determine which genomic regions the reads originated from.
To evaluate the correlation of C1orf132 transcripts originating from the putative p2/distal promoter with the tumor state of breast tissues, we performed data mining of The Cancer Genome Atlas (TCGA) breast cancer subtypes to study their association with the expression level of C1orf132.
Cloning the putative promoters of C1orf132.
The potential promoter regions (p1 and p2) for C1orf132, according to the bioinformatics analysis, were ampli ed from genomic DNA using Accu Taq polymerase kit (Invitrogen, USA) and speci c primers (supplementary Table 1). The ampli ed products (1203 bp for p1 and 2143 bp for p2), were cloned into the promoterless green uorescent protein (pEGFP-1) reporter vector, using anking sequences on primers for AgeI/SalI restriction sites (Supplementary Table 1). The accuracy of the cloned constructs was con rmed by DNA sequencing (Europhins, Italy).
Cell culture and transfection.
We examined the nuclear vs. cytoplasmic subcellular localization of C1orf132 in MCF12A cells by fractionation with cell fractionation buffer (Ambion, USA), according to the manufacturer's instructions. RNA was extracted to assess the relative proportion of C1orf132 in the nuclear and cytoplasmic fractions. The transcription levels of beta 2-microglobulin (B2M) as a cytoplasmic marker, U1 as a nuclear marker and C1orf132 were then assessed using quantitative reverse transcription-polymerase chain reaction (qRT-PCR).
Promoter activity reporter assay.
In order to suppress C1orf132 expression, we decided to delete its p2 promoter using CRISPR/Cas9 system. Three pairs of different guide RNAs (gRNAs) were designed (Supplementary Table 1) to target the putative sequences of the p2 transcription start site, using http://crispr.mit.edu. The three gRNAs where then cloned into TOPO-TA gRNA vector. Brie y, guide RNA plasmid backbone (TOPO-TA gRNA) was linearized using the BbsI enzyme (New England Biolabs, USA) and digested at 37˚C for 1 h. Annealing took place using sense and anti-sense oligonucleotides in buffer 2 (NEB), before placing the reaction in a thermocycler with the ramp of 0.1˚C/sec from 95˚C to 25˚C. To ligate the gRNA within the linearized vector, T4 DNA ligase (Thermo Fisher Scienti c, USA) was used and the reaction was incubated at room temperature for 60 min. Then, E. coli cells were transformed with gRNA vectors using heat-shock protocol followed by plating the cells on LB/ampicillin (100 μg/ml) plates overnight. Colony-PCR was performed to verify the sequence of the gRNA plasmids, using the reverse oligo of guides and forward primer of TOPO vector.
Different pairs of gRNA-containing vectors were co-transfected along with wild type-Cas9 vector (PX458) into MCF12A cells by lipofectamine LTX & PLUS reagent (Invitrogen, USA). Single cells expressing GFP were selected by a cell sorter machine (BD FACSCelesta), 24 h after transfection. Edited colonies were investigated for deletion, by DNA extraction using Puregene Core kit A (Qiagen, USA), according to the manufacturer's instructions, followed by performing PCR using anking primers (Edited-Test) and DNA sequencing (Europhins, Italy).
Wound healing assay.
Edited and unedited MCF12A cells were seeded into 24-well plates (with 6 repeats) and grew to 90% con uency. The monolayers were scratched using a 200 μl pipette tip and then the oating cells were removed by several washes with phosphate buffered saline (PBS). Subsequently, the cells were incubated at 37ºC for 24 h, before being photographed. The migration area of wound healing was border-lined and analyzed, using the ImageJ software.
Cell cycle analysis by ow cytometry.
Triplicates of three different densities (10000, 20000, and 30000) of edited vs. unedited MCF12A cells were seeded into 12-well plates. The Vybrant ™ DyeCycle ™ Violet Stain kit (Invitrogen, USA), which is capable of entering living cells and staining DNA, was used to examine the cell cycle 24 h later. Brie y, cells were washed in cold PBS and then harvested, and resuspended in the complete medium. Next, ow cytometry tubes each containing 1 ml of cell suspension in complete medium at a concentration of 1×10 6 cells/ml were prepared. 1 μl of Vybrant ™ DyeCycle ™ Violet Stain was added to each tube ( nal concentration of 5 μM) and mixed well. After 1 h incubation at 37˚C with protection from light, the samples were analyzed without washing or xing on a ow cytometer (BD FACSCelesta) using laser beam.
RNA extraction and qRT-PCR.
A small amount of frozen patient tissue samples was lysed for RNA extraction using RNSol (ROJE, Iran)/or QIAZOL (Qiagen, USA). Total RNA was extracted according to the manufacturer's instructions. 1 µg of RNA was rst treated with DNase I (Thermo Fisher Scienti c, USA) in order to eliminate any traces of DNA contamination, and then reverse transcribed using PrimeScript rst strand cDNA synthesis kit (TaKaRa, Japan). The expression of target genes was evaluated using BioFACT™ 2X real-time PCR master mix (BioFact, South Korea) through ABI StepOne real-time PCR system. The sequence of primers used for quantifying each target gene can be found in Supplementary Table 1 TaqMan miR assay for miR-29c-3p detection.
Hsa-miR-29c-3p (ID 000587), and U6 snRNA (ID 001973) TaqMan miR assays were ordered from Applied Biosystems (Foster City, CA). Total RNA extraction followed by qRT-PCR assay were performed to determine the expression level of miR-29c-3p in edited versus unedited MCF12A cells, according to the manufacturer's protocol. Brie y, total RNA extraction was performed by Qiazol (Qiagen, USA), followed by reverse transcription using SuperScript II Reverse Transcriptase (Thermo Fisher Scienti c, USA), before doing qPCR using Taqman® Universal PCR Master Mix II (Thermo Fisher Scienti c, USA). The samples were then incubated for 10 min at 95ºC for initial denaturation, and then subjected to 40 PCR cycles, each consisting of 95ºC for 15 sec and 60ºC for 60 sec. U6 was used as the internal control of miR-29c-3p. The 2 −ΔΔCt method was used to analyze microRNA levels.
Statistical analysis.
All statistical analyses were performed using GraphPad Prism 6 software. Student t-test was employed to investigate the signi cance of observed differences of gene expression alterations. All tests were done in 3 biological repeats and values are reported as mean± standard deviation (SD). P values less than 0.05 were considered as statistically signi cant. Figure 1 Genomic location of C1orf132 long-noncoding RNA in human and mouse. Top: The genomic location of miR-29b2, miR-29c and their precursor, MIR29B2CHG, in human. As shown, the region encoding the lncRNA is located between CD34 and CD46 genes. Bottom: By searching the genomic location of miR-29 in mouse, we found out that the lncRNA A330023F24 is the host gene for miR-29b-2 and miR-29c.

Figures
Interestingly the chromosomal location and transcription direction of the lncRNA is conserved between human and mouse.

Figure 2
The genomic features of miR29b2c precursor. To make it easier to follow, the location of miR-29b2 and miR-29c are highlighted with yellow and purple colors, respectively, and the two potential distal   C1orf132 has at least two active promoters. A) The bioinformatics data suggest the existence of two additional potential promoters (p1 and p2), in addition to its conventional promoter. The predicted sequences correlated to the putative promoters were cloned into a promoterless pEGFP-1 vector. MCF7 cells transfected with: B) pEGFP-C1 vector as a positive control, C) pEGFP-1-p2 for the putative p2 promoter region, D) pEGFP-1-p1 for the putative p1 promoter region. Note that only p2 region exhibited promoter activity. E) The subcellular localization of C1orf132 following nuclear and cytoplasmic subcellular fractionation. C1orf132 is primarily localized within the nuclear compartment. U1 and B2M were used as markers for nuclear and cytoplasmic compartments, respectively.   assay was employed as a tool to monitor cell migration in edited (C1orf132-promoter eliminated) vs. unedited MCF12A cells. As it is clearly demonstrated in microscopic pictures, after 24 h the uncovered areas of the plates were signi cantly decreased (p<0.05) in cells edited for C1orf132. E) The expression level of miR-29c-3p in C1orf132-promoter eleminated cells has been signi cantly reduced by half, compared to wild type MCF12A. F) Several genes were differentially expressed in edited vs. unedited