18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License The transcribed ultraconserved region uc.160+ enhances processing and A-to-I editing of the miR-376 cluster: hypermethylation improves glioma prognosis Marta Soler1 , Veronica Davalos1, Anaı́s Sánchez-Castillo2, Carlos Mora-Martinez3, Fernando Setién1, Edilene Siqueira1,4 , Manuel Castro de Moura1, Manel Esteller1,5,6,7 and Sonia Guil1,8 1 Josep Carreras Leukaemia Research Institute (IJC), Barcelona, Spain 2 Department of Radiation Oncology (MAASTRO), GROW School for Oncology and Developmental Biology, Maastricht University Medical Center, The Netherlands 3 Centre of Excellence in Experimental and Computational Developmental Biology, Institute of Biotechnology, University of Helsinki, Finland 4 Conselho Nacional de Desenvolvimento Cientifico e Tecnológico (CNPq), Brasilia, Brazil 5 Centro de Investigación Biomédica en Red Cáncer (CIBERONC), Madrid, Spain 6 Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain 7 Physiological Sciences Department, School of Medicine and Health Sciences, University of Barcelona (UB), Spain 8 Germans Trias i Pujol Health Science Research Institute, Barcelona, Spain Keywords Transcribed ultraconserved regions (T-UCRs) are noncoding RNAs A-to-I editing; glioma; miR-376; noncoding derived from DNA sequences that are entirely conserved across species. RNA; pri-miRNA biogenesis; T-UCR Their expression is altered in many tumor types, and, although a role for T-UCRs as regulators of gene expression has been proposed, their func- Correspondence S. Guil, Josep Carreras Leukaemia Research tions remain largely unknown. Herein, we describe the epigenetic silencing Institute (IJC), 08916, Badalona, Barcelona, of the uc.160+ T-UCR in gliomas and mechanistically define a novel Catalonia, Spain RNA–RNA regulatory network in which uc.160+ modulates the biogenesis E-mail: sguil@carrerasresearch.org of several members of the miR-376 cluster. This includes the positive regu- lation of primary microRNA (pri-miRNA) cleavage and an enhanced A- (Received 11 May 2021, revised 1 October to-I editing on its mature sequence. As a consequence, the expression of 2021, accepted 15 October 2021, available uc.160+ affects the downstream, miR-376-regulated genes, including the online 3 November 2021) transcriptional coregulators RING1 and YY1-binding protein (RYBP) and doi:10.1002/1878-0261.13121 forkhead box P2 (FOXP2). Finally, we elucidate the clinical impact of our findings, showing that hypermethylation of the uc.160+ CpG island is an independent prognostic factor associated with better overall survival in lower-grade gliomas, highlighting the importance of T-UCRs in cancer pathophysiology. Abbreviations ADAR, adenosine deaminase RNA specific; cDNA, complementary DNA; CGI, CpG Island; Chr, chromosome; CI, confidence interval; CRISPR, clustered regularly interspaced short palindromic repeats; EMSA, electrophoretic mobility shift assay; EV, empty vector; FOXP2, forkhead box P2; GBM, glioblastoma multiforme; Hsa, homo sapiens; Hg, human reference genome; HR, hazard ratio; HRP, horseradish peroxidase; IDH, isocitrate dehydrogenase; LGG, low-grade glioma; miRNA/miR, microRNA; Mut, mutant; Ns, not significant; OS, overall survival; PCR, polymerase chain reaction; pre-miR/pre-miRNA, precursor microRNA; pri-miR/pri-miRNA, primary microRNA; RT, retrotranscription; RT-qPCR, retrotranscription quantitative PCR; RYBP, RING1 and YY1-binding protein; SCR, scramble; SD, standard deviation; sg/sgRNA, single guide/single guide RNA; TCGA, The Cancer Genome Atlas Program; T-UCR, transcribed ultraconserved region; UCR, ultraconserved region; UCSC, University of California Santa Cruz; UTR, untranslated region; WT, wild-type; χ2, chi-squared test. 648 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License M. Soler et al. T-UCR uc.160+ regulates miR-376 in glioma We have previously described how miRNA biogene- 1. Introduction sis can be regulated by T-UCRs: uc.283+ controls pri- Over the last decade, noncoding RNAs have been miRNA processing through RNA:RNA complemen- shown to play a variety of regulatory roles in gene tarity with the lower stem region of the pri-miR-195, expression networks, and thereby to have a broad impairing miRNA biogenesis at the level of Drosha influence on physiopathology [1]. Despite their hetero- cleavage [16]. The biogenesis of miRNAs is a highly geneous origin and structure, one common feature is regulated process that can affect the amount and iden- their relatively poor conservation across species [2]. tity of mature miRNA [22]. Also, changes in the criti- However, there are some striking exceptions: The cal seed region of a miRNA (nucleotides 2–8 from the human genome contains several hundred ultracon- 50 end of the miRNA) [23,24] can reassign its speci- served regions (UCRs) that are 100% identical in ficity for target mRNAs. In the work reported here, human, mouse, and rat genomes, and are extensively we have identified uc.160+ as an epigenetically regu- transcribed, producing a class of long noncoding lated T-UCR in human gliomas. Mechanistically, RNAs known as T-UCRs [3,4]. Most of these are uc.160+ regulates miR-376 cluster biogenesis through expressed in a tissue-specific manner in normal cells. complementarity with the lower stem sequence of the Their functional relevance became evident with the pri-miRNAs. This promotes Drosha cleavage and A- identification of deregulated T-UCR signatures associ- to-I editing of the mature miRNAs, with consequences ated with specific disease conditions [4-7]. Moreover, for the regulated mRNA targets. altered profiles of T-UCRs have been proposed as prognostic factors in human malignancies [8,9]. 2. Materials and methods Similar to other long noncoding RNAs, T-UCRs may influence tumorigenesis by participating in proliferation, apoptosis, migration, or invasion, but little is known about 2.1. Cell culture their mode of action. For example, the uc.454 T-UCR increases apoptosis in lung cancer through direct interac- Human glioma cell lines U-87 MG and KS-1 (pur- tion with the 3’UTR of HSPA12B mRNA [10]. Uc.338 chased from the Japanese Collection of Research inhibits p21 signaling by interacting with the BMI1 poly- Bioresources Cell Bank) were maintained in Dul- comb family member [11], modulating the PI3K/AKT becco’s modified Eagle’s medium (DMEM), supple- pathway [12], and negatively regulating TIMP-1 30 UTR mented with 10% v/v fetal bovine serum (FBS, Gibco, [13]. Importantly, many T-UCRs are suspected of affecting Waltham, MA, USA) at 37 °C in a humidified atmo- miRNA abundance and function, mainly as a consequence sphere of 5% CO2 and 95% air. Normal human astro- of sequence complementarity between the two classes of cytes were purchased from Innoprot (#P10251). All RNAs [14,15]. In a few cases, T-UCRs have been shown cell lines were routinely checked for mycoplasma con- to interfere with miRNA biogenesis through base-pairing tamination. with the primary miRNA transcript [16-18], and likewise, T-UCRs can be targeted by miRNAs. For example, over- 2.2. Search for matches between pri-microRNAs expression of miR-155 in leukemia cells reduces the levels and T-UCRs of uc.160+ [4,14,19], and miR-153 suppresses uc.416 expression in gastric cancer [19]. Based on microarray Ultraconserved element sequences were downloaded data, it was proposed that the T-UCR signature in neu- from UCbase (www.ucbase.unimore.it) [25]; the strand roblastoma prognosis groups is at least partially explained from which they are transcribed is provided in ref. [4]. by the miRNA profile [14]. T-UCR expression can also be microRNA hairpin sequences were obtained from MIR- modulated by epigenetic mechanisms. For instance, BASE v.20 (www.mirbase.org) [26]. To find the matches changes in local DNA methylation are associated with between pairs of sequences, a regex-based algorithm dysregulation of T-UCRs in a variety of tumor types, and was implemented in Perl. The program compared in some cases, are correlated with tumor stage. Hyperme- every substring of a given minimum length (which was thylation of uc.283+A, uc.160+, and uc.346+ CpG island set to 11) from one database to another. When a is associated with silencing in cancer cells [7,20]. Moreover, match was found, adjacent nucleotides of the two the cancer specificity of these methylation events highlights sequences were sequentially compared with elongate it. their potential as noninvasive biomarkers in circulating Identical sequence matches between different hairpins DNA, as recently shown in plasma samples of colorectal of the same microRNA cluster and a T-UCR were cancer patients [21]. grouped accordingly. Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of 649 Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T-UCR uc.160+ regulates miR-376 in glioma M. Soler et al. (Qiagen, Hilden, Germany), according to the manufac- 2.3. Plasmid construction and generation of turer’s protocol. Cells were harvested 48 h after trans- mutants by direct mutagenesis fection and analyzed by western blot. The ultraconserved region within uc.160+ was cloned from DNA into the pcDNA3.1(+) vector (Invitrogen), 2.5. miR-376 cluster editing by CRISPR/Cas9 using primers that introduced the BamHI and EcoRV restriction sites for directional cloning. Pri-miRNA CRISPR/Cas9 technology was used for engineering hsa- sequences, including ˜ 150 bp upstream and down- miR-376 cluster knockout in the KS-1 cell line. Two stream of the hairpins, were cloned from DNA into sgRNAs (sgRNA1: 50 -GCACTTTGCGAGTCCCA the pSPARK® TA vector (Canvax Biotech, Cordoba, CGT-30 and sgRNA2: 50 -ATGGTGAGAGCAGCACA Spain), with the sense orientation under T7 promoter, CCG-30 ) were designed using the CRISPR Design tool and expanded in E. coli DH5α bacteria. To generate page (http://crispr.mit.edu/) and cloned into the BbsI uc.160+ and pri-miRNA mutants, overlapping primers sites of pSpCas9 (BB)-2A-GFP (PX458) (#48138, that introduced the desired mutation were designed Addgene®, Teddington, UK). Cells in a 100-mm culture and used in PCRs with AccuPrimeTM Pfx DNA Poly- dish were transfected at 70% confluence with 8 µg (4 µg merase (Thermo Fisher Scientific, Waltham, MA, sgRNA1 and 4 µg sgRNA2) of construct plasmids with USA) to amplify the wild-type plasmids. PCRs were LipofectamineTM Stem Transfection Reagent (Thermo then treated with 1 U of DpnI restriction enzyme Fisher Scientific). Two days after transfection, cells were (Takara) for 1.5 h at 37 °C to remove the parental detached and resuspended in 1X PBS, 2 mM EDTA, plasmids. All oligos used are listed in Table S1. and 0.5% FBS, for flow cytometer selection. Cells con- taining green fluorescence (eGFP+) were selected by FACS and pooled. End-point PCR was used to confirm 2.4. Transient transfections the deletion of the target region. PCR primers are listed U-87 MG and KS-1 cells were transfected at 75–80% in Table S1. confluence with 8 µg of construct plasmids (pcDNA3.1- uc.160+, pcDNA3.1-uc.160+ mut12, or empty pcDNA3.1 2.6. Nuclear and cytoplasmic fractionation vector as negative control) in a 100-mm culture dish with jetPRIME® transfection reagent according to the manu- Subcellular fractionation was performed with a PAR- facturer’s recommendations (1 : 2 DNA to jetPRIME® ISTM kit (#AM1921, Life Technologies) as previously ratio (w/v)). Cells were harvested 48 h after transfection. described [27]. Equal amounts of RNA from each frac- Synthetic mimics of hsa-miR-376a-3p tion were subjected to RT-qPCR, and the results were (MIMAT0000729), hsa-miR-376c-3p (MIMAT0000720), calculated using the comparative Ct method 2ðΔCt Þ and a negative control miRNA (scrambled, scr) were and shown as a percentage, considering the total quan- purchased from Shanghai GenePharma (Shanghai, tity of RNA recovered from each fraction. To verify China). The edited forms of hsa-miR-376a-3p and hsa- the nuclear and cytoplasmic fractionation of the miR-376c-3p were purchased from Sigma-Aldrich mRNA, RNU6B and GAPDH were used as controls, (St.Louis, MO, USA). Cells were plated onto a 100-mm respectively. The separation was confirmed at the pro- dish and transfected at 30–50% confluence with 40 nM tein level by western blot with HISTONE H3 of each miRNA mimic by using Lipofectamine 2000 (#ab1791, Abcam, Cambridge, UK, 1 : 5000) and α- transfection reagent (Thermo Fisher Scientific), accord- TUBULIN HRP (#ab40742, Abcam, 1 : 5000). ing to the manufacturer’s protocols. Cells were har- vested 48 h after transfection and analyzed in the 2.7. RNA extraction and RT-qPCR following assays: RT-qPCR, western blot, and editing. The oligonucleotides used are listed in Table S1. Total RNA, including miRNAs, was extracted with a MiRIDIAN® microRNA Hairpin Inhibitors (an- Promega Maxwell® RSC miRNA Tissue kit (AS1460, tagomiRs) against hsa-miR-376a (IH-300683-05-0005), Promega, Madison, WI, USA) according to the manu- hsa-miR-376c-3p (IH-300674-06-0005), and a control facturer’s recommendations. For expression analysis, scrambled antagomiR (miRIDIAN microRNA Hairpin total RNA was reverse-transcribed using the Rev- Inhibitor Negative Control (IN-001005-01-05)) were ertAid H minus Reverse Transcription Kit (EP0451, purchased from Dharmacon (Lafayette, CO, USA). Thermo Fisher Scientific) with either oligodT primer AntagomiR transfections were performed at a final con- (for mRNAs) or random primers (for T-UCR). A neg- centration of 200 nM (100 nM of Ant-376a and 100 nM ative control minus reverse transcriptase was run in of Ant-376c), using HiPerFect Transfection Reagent parallel to control for genomic contamination. Real- 650 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License M. Soler et al. T-UCR uc.160+ regulates miR-376 in glioma time PCRs were performed in triplicate in a QuantStu- GE HealthCare, 1 : 5000). ECL reagents (Luminata- dioTM 5 Real-Time PCR system (Thermo Fisher Scien- HRT, Merck-Millipore, Burlington, MA, USA, and tific), using 30–100 ng cDNA, 6 µL SYBR® Green SuperSignal West Femto, Thermo Fisher Scientific) and PCR Master Mix (Thermo Fisher Scientific), and the iBrightTM CL1500 Imaging System (Thermo Fisher 416 nM primers in a final volume of 12 µL for 384-well Scientific) were used to visualize the proteins. plates. All data were acquired and analyzed with QuantStudioTM Design & Analysis Software v1.3.1 and 2.9. In vitro pri-miRNA processing assay normalized with respect to the endogenous controls, GUSB, PPIA, and HPRT1. Relative RNA levels were Pri-miR-376 RNA substrates for in vitro processing calculated using the comparative Ct method 2ðΔΔCt Þ . assays were prepared from DNA templates by stan- For miRNA expression analysis, the miRCURY dard in vitro transcription with T7 RNA Polymerase LNATM miRNA PCR assay system (Qiagen) was used, (Roche) in the presence of [α-32P]-ATP (PerkinElmer), following the manufacturer’s recommendations, with as previously described [16]. RNA substrates corre- the miRCURY LNA RT kit (Cat. No. 339340, Qia- sponding to the ultraconserved region of uc.160+ (322 gen) for RNA retrotranscription, and the miRCURY nucleotides) were obtained by in vitro transcription LNA SYBR® Green PCR Kits (Cat. No. 339345, Qia- from linearized DNA templates. Processing reactions gen) for the RT-qPCR, in a QuantStudioTM 5 Real- were carried out with total HEK293T extracts, as pre- Time PCR (Thermo Fisher Scientific) apparatus with viously described [28]. Briefly, each pri-miRNA was QuantStudioTM 5 software. To normalize the data, incubated with increasing amounts (0.071, 0.155, and RNU6B, miR-191-5p, and miR-423-3p were used as the 0.284 μM) of uc.160+. The RNA mixture was pre- endogenous controls. In vivo pri-miRNA processing heated at 65 °C for 2 min and then cooled to 30 °C. was performed following total RNA extraction and The nuclear extract was added, and the reaction mix- reverse transcription with random primers and Super- ture incubated for a further 90 min at 30 °C. RNAs scriptTM III Reverse Transcriptase (Cat. No. 18080044, were phenol-extracted, precipitated, and loaded onto Thermo Fisher Scientific). Oligos used for qPCR are an 8% denaturing polyacrylamide gel. listed in Table S1. 2.10. Determination of pri-miR-376 cluster RNA 2.8. Western blot editing sites Cell pellets were resuspended in Laemmli SDS sample To measure the editing levels of hsa-miR-376 family buffer (10% glycerol, 2% SDS w/v, 63 mM Tris/HCl pH members, we followed the protocol previously 6.8, 0.01% bromophenol blue) plus 2% 2- described [29]. Total RNA was extracted from cell pel- mercaptoethanol, sonicated, and boiled for 5 min. Equal lets using TRIzol® reagent (Invitrogen) and treated amounts of protein extracts were loaded onto Tris- twice with DNase (#M6101, RQ1 RNase-Free DNase, Glycine-SDS gels and transferred to a nitrocellulose Promega). First-strand cDNA was synthesized from membrane (Whatman, GE Healthcare, Chicago, IL, 4 µg of total RNA with the SuperScriptTM III Reverse USA), by liquid electroblotting (Mini Trans-Blot Cell, Transcriptase (Cat. No. 18080044, Thermo Fisher Sci- Bio-Rad, Hercules, CA, USA) for 1 h at 100 V. Mem- entific), using two pri-miRNA-specific RT primers: the branes were blocked and incubated overnight at 4 °C ‘pri-miR376a2-c editing Rv’ primer for pri-miRNA- with primary antibodies diluted in 5% nonfat milk in 376a2 cluster and the ‘pri-miR376a1-b editing Rv’ pri- PBS containing 0.1% Tween-20. The proteins detected mer for the pri-miRNA-376a1-b cluster (Table S1). As were as follows: RYBP (#ab185971, Abcam, 1 : 1000), a negative control, the same reactions were carried out FOXP2 (#5335, Cell Signaling Technology, Denvers, with 4 µg of RNA without reverse transcriptase MA, USA, 1 : 1000), LAMIN B1 (#ab16048, Abcam, enzyme. cDNA products were then amplified with 1 : 5000), α-TUBULIN HRP (#ab40742, Abcam, Immolase Taq polymerase (Bioline), using the specific 1 : 5000), HISTONE H3 (#ab1791, Abcam, 1 : 5000), forward and reverse PCR primers for each pri-miRNA ADAR1 (AMAB90535, Atlas Antibodies, 1 : 1000), and cDNA (Table S1). Products were isolated from agar- ADAR2 (HPA018277, Atlas Antibodies, Bromma, Swe- ose gel bands using a NucleoSpin® Gel and PCR den, 1 : 400). After three washes with PBS containing Clean-up (Macherey-Nagel, Düren, Germany) and 0.1% Tween-20, membranes were incubated for 1 h at sequenced in a 3730 DNA Analyzer (Applied Biosys- RT in a bench-top shaker with the secondary antibodies tems). All RT-PCR products were subcloned into the conjugated to horseradish peroxidase anti-rabbit IgG pGEM®-T Easy Vector Systems (Promega), following (A0545, Sigma, 1 : 10 000) or anti-mouse IgG (Na9310V, the manufacturer’s protocol. For each cDNA, 75–100 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of 651 Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T-UCR uc.160+ regulates miR-376 in glioma M. Soler et al. clones were isolated and sequenced. Following analysis proportional hazards regression was used to identify the with BIOEDIT v7.2.5 software, the frequency of editing independent variables associated with OS. Experimental was quantified as the ratio of the number of A-to-G groups were compared using Student’s unpaired-sample changes to the total number of cDNA clones test. For association between variables, such as editing sequenced. A-to-I editing sites are indicated with a G frequency, we used chi-squared contingency and Fish- in the chromatogram. er’s exact tests. All statistical tests were two-sided. Levels of significance were recognized as P < 0.05 (*), < 0.01 (**) and < 0.001 (***). 2.11. Electrophoretic mobility shift assay (EMSA) The RNA substrates for uc.160+, uc.160+ mut5, and 3. Results and Discussion uc.160+ mut12 were obtained and biotin-labeled during the in vitro transcription by using 0.25 mM biotin-16- UTP (Roche) in the transcription reaction. Wild-type 3.1. The ultraconserved transcript uc.160+ is or mut5 pri-miRNA substrates were also synthesized by commonly methylated in glioma samples and in vitro transcription from linearized DNA templates. cell lines Binding reactions were carried out in 1X binding buffer (20 mM Tris/HCl pH 8.0, 1 mM DTT, 1 mM MgCl2, We have previously described that uc.160+ undergoes 20 mM KCl, 10 mM Na2HPO4-NaH2PO4 pH 8.0) with cancer-specific hypermethylation-associated transcrip- the biotin-labeled RNA alone (0.1 pmol) or in the pres- tional silencing in some tumor types [20]. The full- ence of increasing amounts (0.5–4 pmol) of unlabeled length transcript has been catalogued in the MiTran- T-UCR or pri-miRNAs, in a final volume of 15 µL. scriptome database of long polyadenylated RNA tran- Each RNA mixture was preheated at 70 °C for 5 min, scripts (www.mitranscriptome.org) with reference gradually cooled down to let the RNA regain its native G066395|T284682, although it has not been yet anno- structure, and then left at 30 °C for 20 min. All reac- tated in Refseq (Fig. 1A). We first assessed the cancer- tions were then immediately loaded onto a native 6% specific hypermethylation of uc.160+ by surveying nor- polyacrylamide gel (29 : 1 acrylamide:bis-acrylamide), mal and tumor samples from The Cancer Genome transferred to a nylon membrane, and developed using Atlas (TCGA) datasets, which revealed that > 50% of a BrightStar® BioDetectTM Nonisotopic Detection Kit cases of most common primary tumor types (including system (Thermo Fisher Scientific). gliomas, breast, and colon cancer) are hypermethylated (Fig. 1B, Table S2). By contrast, normal tissues show low or no methylation (Fig. S1A, Table S2). 2.12. Databases and statistical evaluation We focused our study in gliomas, the most common, miRBase (http://www.mirbase.org) and TargetS- and lethal primary intracranial tumors. In accordance canHuman (http://www.targetScan.org) were used to with the primary tumors, uc.160+ was commonly predict binding sites for the miR-376 family, and hypermethylated in a panel of glioma cell lines but not miRDB (http://mirdb.org) for target custom prediction in normal astrocytes (Table S2). Importantly, tran- of the edited miRNA forms. DNA methylation data of scriptional silencing was observed in the hypermethy- lower-grade glioma (LGG) and glioblastoma multiforme lated cell lines, confirming the previously reported (GBM) cases were extracted from the TCGA Data Por- epigenetic regulation of uc.160+ [20] (Fig. 1C). Analy- tal (https://portal.gdc.cancer.gov/). The glioma cohort sis of U-87 MG cDNA by RT-PCR detected the included patients for which methylation on the uc.160+ MiTranscriptome-annotated transcript, mostly in the locus, IDH1 mutation, and survival was available. unspliced form (Fig. S1B). In addition, cellular frac- Methylation levels (β-value) of CpG sites > 0.33 were tionation showed that uc.160+ was enriched in the considered to be hypermethylated. Graphs and statisti- nucleus, as is the case for many unspliced transcripts cal comparisons were obtained with the GRAPHPAD PRISM that are deficiently exported (Fig. 1D), and in silico 9.0.0 (La Jolla, CA, USA) and IBM SPSS Statistics exploration of its coding potential drew attention to (Armonk, NY, USA) for Windows. We used the the transcript’s noncoding nature (Fig. S1C). Kaplan–Meier method to analyze patient data to esti- mate survival. The log-rank test was used to establish 3.2. uc.160+ is complementary to the miR-376 any differences between patient groups. Hazard ratios cluster and enhances its processing (HRs) from univariate Cox regressions were used to determine the association between clinicopathological Since our previous work and that of others have features and overall survival (OS). Multivariate Cox shown that some T-UCRs regulate miRNA biogenesis 652 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License M. Soler et al. T-UCR uc.160+ regulates miR-376 in glioma A C chr5 (hg19) U-87 MG Relation_to_ LN229 77,268,000 77,268,500 77,269,000 77,269,500 77,270,000 77,270,500 77,271,000 Map Info A172 Cas1 KS-1 UCSC_CpG_ Probe ID G066395|T284682 (GRCh38/hg19) Island UCR CGI chr5: 77268452 cg04502985 Island 0.12 0.97 0.98 0.97 0.99 chr5: 77268624 cg11847222 Island 0.2 0.9 0.88 0.88 0.9 376 637 326 165 26 80 530 419 87 10 593 124 156 chr5: 77268734 cg13031251 Island 0.16 0.97 0.96 0.96 0.97 B n= 413 429 748 257 838 292 133 439 264 172 57 222 508 100 TCGA primary tumors (%) uc.160+ hypermethylated 0.015 uc.160+ Relative expression (2- Ct) 80 60 0.010 40 20 0.005 0 t Lu ct St ost as Es B ace T r ix Eyg U Pa ne er Ly y s yr is op re h Enrina ncr ck M om y t a s G sti in re th riumt tis ry So Oland Th est d Pr iom e lia Cegus l g ium an L e d d e d Th teruy ph s ha as Ad eso et rac U ne m mu om a t n T oi l n oi Ki su ry rv a te k ft va iv d r e in S na e l d 0.000 e d Bi rg ea an La H 72 1 9 -1 ic G as 22 et KS M A1 oi C LN op 7 at -8 em U H D E uc.160+ T-UCR miRNA T-UCR Match sequence Match Hairpin U-87 MG KS-1 position length region 100 N C N C HISTONE H3 hsa-miR-1289-2 245 UCCUAGUUAAAAA 13 ss-ds hsa-miR-376a-1, a-2, b, c 184 AGUAUCAAAUGC 12 ss-ds % total -TUBULIN 50 hsa-miR-1248 175 UGCAAUAAUGCU 12 ss-ds hsa-miR-5004 293 CUUCCUUUCAGA 12 ds-ss uc.160+ hsa-miR-124-1 75 GAUUUAAAUGU 11 T. Loop hsa-miR-3202-1 20 GGAGUCAUUUU 11 T. Loop 0 hsa-miR-4711 70 AAAUGUGCAUC 11 ds-Mature N C N C hsa-miR-577 59 UCAUUUCCCCA 11 ss-ds hsa-miR-640 268 UCCCUUAUCUG 11 ss-ds U-87 MG KS-1 CNNC box pri-miR-376c pri-miR-376a-1 pri-miR-376a-2 pri-miR-376b mature 5p mature 3p complementary to uc.160+ F G pri-miR-376c U U U Pri-miR-376c UGGUA UGUUA uc.160 _ UUCCU CUAUGUUUA AAGGA GAUACAAAU _ _ _ _ _ _ _ _ pri-miRNA U _ - _ _ _ _ 150- UA UA _ AAAA GUGGA UGA UUUU CACCU _ _ _ _ 100- G G _ _ _ _ 90- 80- U C AUU 70- pre-miRNA UC GGACUCAAUCCUUCUUUGGU AUCAAAUGCUGCUUGGAAACAUU pri-miR-376a-1 pri-miR-376b pri-miR-376c pri-miR-376a-2 60- A UAGUUUACG 50- uc.160 40- uc.160+/pri-miRNA duplex 30- Biotin-uc.160+ through complementarity with their primary nucleotides) was found with pri-miR-1289, which is sequences, we next looked for potential complementar- expressed at a very low level (mirbase.org, release ity (of at least 11 nucleotides) between uc.160+ and 22.1) and for which no role in brain pathophysiology miRNA primary sequences, excluding the mature has been identified. Hits with 12 nucleotides included regions (Fig. 1E). The longest complementarity (13 miR-1248 (also expressed at a very low level, according Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of 653 Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T-UCR uc.160+ regulates miR-376 in glioma M. Soler et al. Fig. 1. Methylation of the ultraconserved transcript uc.160+ and complementarity with miR-376 family. (A) uc.160+ genomic region and associated CpG island. Transcription of the RNA including the uc.160+ ultraconserved region is supported by the annotated transcript G066395|T284682 in the MiTranscriptome database (mitranscriptome.org). Uc.160+ coordinates are those given in the GRCh37/hg19 release. The black bar shows the sequence of uc.160+ from the UCbase 2.0. database. The green bar shows the associated CpG island (CGI). (B) Percentage of uc.160+ methylation in the TCGA panel of samples by tumor type. (C) Above, DNA methylation profile of the uc.160+-associated CpG island in five glioma cell lines, analyzed by the 450K DNA methylation array in a previous study [69]. Single CpG absolute methylation levels (0–1) are shown. Green, unmethylated; red, methylated. Below, expression levels of uc.160+ in the same cell lines as determined by real-time PCR (n = 1). (D) Nuclear/cytoplasmic fractionation of U-87 MG and KS-1 cell lines, analyzed by RT-qPCR and western blot to assess fraction purity. Graphs represent the mean  SD of n = 2 replicates of fractionation. (E) Table above, Complementarity between uc.160+ and pri-miRNA hairpins, listed by match length. Only matches ≥ 11 nucleotides are shown; the region of complementarity is indicated. Below, hairpin regions of all pri-miRNAs of the miR-376 cluster are aligned, and the region of complementarity to uc.160+ is indicated. Results obtained with BIOEDIT v7.0.5.3. (F) Above, diagram of the mir-376 cluster; nomenclature is from the miRBASE (http://microrna.sanger.ac.uk/sequences). Numbers at the bottom indicate intervening lengths of sequences. Created with Biorender.com. Below, electrophoretic mobility shift assay (EMSA) with biotin-labeled uc.160+ RNA and increasing levels of unlabeled pri-miRNA-376a-1/b/c or a-2. (0.5-1-2-4 pmols). (G) Left, diagram to illustrate the region of complementarity (highlighted in green) between uc.160+ and pri-mir-376c at the base of the stem, near the CNNC box (red box). The mature miRNA sequence is highlighted in blue. Hairpin structure is depicted according to [32]. Black arrows indicate Drosha cropping sites. Right, in vitro pri-miRNA processing assays with 32P- labeled pri-miR-376c in the presence of increasing amounts of uc.160+. to the TissueAtlas, https://ccb-web.cs.uni-saarland.de/ regulated, we next investigated whether this binding tissueatlas/ [30]), miR-5004 (expressed at a very low affected pri-miRNA cleavage. Many protein factors are level and very poorly characterized), and the miR-376 known to regulate this [33], but to our knowledge, only family, which has the highest level of expression in the very few RNAs directly control miRNA biogenesis brain [30] (Fig. S2A), and whose lower level of expres- [16,34,35]. We ran in vitro processing assays with labeled sion in glioma predicts poor outcome [31]. These facts pri-miRNA sequences and confirmed that addition of prompted us to investigate a possible interplay the ultraconserved region of uc.160+ enhanced cleavage between uc.160+ and the miR-376 family in gliomas. and release of the pre-miRNA. This is prominent with This family is transcribed as a cluster of pri-miRNAs pri-miR-376c, which, in accordance with the high levels that include pri-miR-376c, pri-miR-376a-2, pri-miR-376b, of miR-376c found in vivo, is very efficiently processed in and pri-miR-376a-1, from which the three most common vitro (Fig. 1G), and is also seen with the other pri- mature miR-376s are produced (all from 3p arms): miR- miRNAs from the miR-376 family (Fig. S2F–G). This 376c (the most abundant), miR-376a, and miR-376b regulation, together with the proximity of the CNNC (Fig. 1E–F, Fig. S2A–B). All four pri-miR-376 hairpins box suggests that several factors, including ncRNAs, have the 12-nucleotide site complementary to uc.160+ at may converge around this 3’ flanking site to influence the base of the stem-loop structure, at the junction with pri-miRNA cleavage. the single-stranded flanks (Fig. 1E–F). This region is within a stretch of 21 nucleotides that is identical in all 3.3. uc.160+ enhances A-to I editing of miR-376 pri-miRNAs of the family, is even more conserved than family members the mature 5p or 3p miRNAs (Fig. 1E), and resides next to a ‘CNNC’ box (one structural motif on pri-miRNAs We next aimed to measure the impact of uc.160+ on that allows accurate processing through the recruitment pri-miR-376 processing in cells. In all glioma cell lines of auxiliary factors [32,33]). Direct binding between the analyzed, endogenous uc.160+ is expressed at low pri-miRNA sequences and uc.160+ was confirmed by in levels, so we assayed the impact of its ectopic expres- vitro binding assays (Fig. 1F). We then generated the sion. Overexpression in U-87 MG and KS-1 glioma uc.160+ mut12 and uc.160+ mut5 substrates, in which 12 cell lines resulted in a 2- to 3-fold increase in the levels or 5 of the complementary nucleotides were mutated. In of mature miR-376a, b, and c, confirming the ability of addition, compensatory mutations on each pri-miRNA the T-UCR to enhance miR-376 cluster processing were also introduced to match uc.160+ mut5 (Fig. S2C) (Fig. 2A–B). Moreover, this is accompanied by a con- (uc.160+ mut12 could not be fully compensated on the comitant reduction in pri-miR-376c levels (the only pri- pri-miRNAs without potentially disrupting their hairpin miRNA we could robustly detect by RT-qPCR) and is structures). The use of these mutants in in vitro binding abolished when the uc.160+ mut12 was used, confirm- assays indicated that the complementary site was ing that the complementary region is required for reg- required for the interaction (Fig. S2D–E). Since miRNA ulation (Fig. 2A–B). This prompted us to further biogenesis at the level of Drosha processing is tightly analyze miR-376 biogenesis: In the brain, members of 654 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License M. Soler et al. T-UCR uc.160+ regulates miR-376 in glioma the miR-376 cluster are modified by hydrolytic deami- sites), and pri-miR-376b (+67 site), as measured by nation of adenosine to inosine (A-to-I editing) [36]. A- cDNA cloning and sequencing (Fig. 2C). Since miR- to-I editing is catalyzed by adenosine deaminases 376b is expressed at very low levels in U-87 MG (Fig. (ADARs), which require dsRNA for binding and edit- S2B), we focused on pri-miR-376c and pri-miR-376a-2 ing, and is an important source of transcriptomic and confirmed the hyperediting upon uc.160+ overex- divergence from genomic DNA [37]. Inosines are func- pression by directly sequencing the cDNA (Fig. 2D). tionally equivalent to guanosines, and therefore, A-to-I Position +11 on pri-miR-376a-2 is only marginally editing can directly affect the amino acid sequence of hyperedited and falls outside the mature 5p miRNA, certain proteins when located on mRNA codons, as is and so was not further explored. the case for several neurotransmitter receptors and ion Conceptually, one way of regulating A-to-I editing is channels [38], thereby modulating neuronal signaling to alter the abundance of ADARs, but analysis by west- [39]. A-to I editing is tightly regulated, and disruption ern blot revealed no changes in ADAR1 and ADAR2 of this process associates with neurological disorders protein levels upon T-UCR overexpression, with varying and some types of cancer [40]. The systematic analysis levels of these enzymes being present in a panel of of A-to-I editing in the TCGA datasets indicates that glioma cell lines (Fig. S3A–B). Rather, by analogy with some nonsynonymous RNA editing events may be clini- how RNA-binding proteins can alter editing in a site- cally relevant master driver events with crucial roles in specific manner (e.g., by changing the dsRNA structure cancer [41]. Most A-to-I substitutions are found on non- of the target [52,53]), we can hypothesize that base- coding transcripts [42,43], of which miRNAs are the pairing between uc.160+ and pri-miRNA transcripts can best studied examples [44-46]. When A-to-I editing influence editing efficiency by promoting structural affects the seed region of a particular miRNA, it can changes in the pri-miRNA stem-loop that enhances modify target specificity and alter the profile of regu- ADAR binding. Alternatively, this could be mediated lated mRNAs [47,48]. miRNA editing can occur at the by the Microprocessor itself, which acts as a recruiter of pri-miRNA and pre-miRNA levels, sometimes prevent- ADAR enzymes toward the target. In this second sce- ing the miRNA from maturing [49], or from loading nario, RNA:RNA interactions between T-UCR and onto RISC [50], or even leading to the degradation of miRNA would favor Microprocessor recognition and, the miRNA [51]. By contrast, previous studies have indi- concomitantly, the recruitment of editing enzymes. cated that, in the case of miR-376 RNAs, editing does not affect the efficiency of the processing steps [36]. 3.4. Identification of RYBP and FOXP2 as Given the link between editing and miRNA biogenesis, downstream targets we next investigated whether the levels of editing of the miR-376 cluster were altered by uc.160+ overexpression. Previous studies have reported the impact of mature Enforced expression of uc.160+ (but not of uc.160+ miR-376a/c-5p editing on mRNA target regulation mut12) in U-87 MG cells significantly increased editing [36,54], but editing on the 3p arms, which, according of pri-miR-376c (+48 site), pri-miR-376a-2 (+11 and +55 to miRBase, gives rise to the most strongly expressed Fig. 2. Influence of uc.160+ on the processing and editing of the miR-376 family in glioma cell lines. (A, B) RT-qPCR analysis of the endogenous levels of pri-miR-376c or mature miR-376a, miR-376b, and miR-376c in U-87 MG (A) and KS-1 cell lines (B) transiently transfected with uc.160+, uc.160+ mut12, or empty vector (EV). Graphs represent the mean  SD of three (A) or four (B) independent RNA extractions. Two-tailed Student unpaired-sample t-tests were used to evaluate group differences (*P < 0.05, **P < 0.01, ***P < 0.001, ns = not significant). (C) Quantification of editing frequencies for miR-376 cluster editing sites in U-87 MG cell line upon transient transfection of uc.160+ or uc.160+ mut12 vectors (empty vector was used as control). Sanger sequencing of RT-PCR clones of primary miRNAs (pri-miRs) from three independent experiments was employed. The increase in RNA editing in pri-miR-376c (position +48) and pri- miR-376a-2 (position +55) upon uc.160+ overexpression is highlighted in gray. For each pri-miRNA, the 50 end of the stem-loop sequence annotated in the Sanger miRBase dataset is counted as +1, as illustrated in the upper diagram of the cluster (created with Biorender.com). Editing frequency is calculated as the ratio of the number of A-to-G changes to the total number of cDNA clones sequenced. The statistical significance of the difference between empty vector and uc.160+ or uc.160+ mut12 overexpressed samples was assessed by chi-squared contingency and Fisher’s exact tests. (D) Left panel: chromatograms of direct Sanger sequencing of RT-PCR products corresponding to pri- miR-376c and pri-miR-376a-2 upon transient overexpression of uc.160+ or uc.160+ mut12. A-to-I editing is detected as an A (green) to G (black) peak in the cDNA sequence. Black arrows indicate potential editing sites, and the most highly edited positions are boxed. Right panel: contingency graphs of the editing frequency of pri-miR-376c (position +48) and pri-miR-376a-2 (position +55), as identified in (C). The y-axis shows the number of wild-type (WT) and edited clones, as a percentage. Editing frequency was assessed by chi-squared contingency and Fisher’s exact tests (*P < 0.05; **P < 0.01). Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of 655 Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T-UCR uc.160+ regulates miR-376 in glioma M. Soler et al. U-87 MG U-87 MG KS-1 A B miR-376a 4 miR-376a miR-376b 4 miR-376b * Relative expression pri-miR-376c Relative expression 2.0 miR-376c Relative expression miR-376c ** *** * 3 *** 1.5 *** 3 ** 1.0 2 ns 2 ns * 0.5 1 1 0.0 0 0 .1 c. V + 0+ 12 EV + 12 EV + 12 uc u E 60 60 ut 60 16 ut ut .1 .1 m m m uc uc + + 60 60 .1 .1 uc uc C # edited sites pri-miR-376c pri-miR-376a-2 pri-miR-376b pri-miR-376a-1 Site (+4) (+48) (+11) (+15) (+55) (+67) (+5) (+9) (+49) empty vector 0/94 (0 %) 19/92 (20.6 %) 0/94 (0 %) 4/94 (4.3 %) 43/95 (45.3 %) 46/75 (61.3 %) 3/75 (4 %) 0/85 (6.3 %) 6/75 (8 %) uc.160+ 2/81 (2.4 %) 32/81 (39.5 %) 7/81 (8.6 %) 6/81 (7.4 %) 51/81 (62.9 %) 68/85 (80 %) 0/85 (0%) 3/75 (4 %) 14/85 (16.4 %) uc.160+mut12 1/94 (1 %) 18/94 (19.1 %) 4/94 (4.3 %) 0/94 (0 %) 50/94 (53.2 %) 26/39 (66.6 %) 0/39 (0 %) 0/39 (0 %) 3/39 (7.7 %) uc.160+ 0.2128 0.0077 0.0039 0.5167 0.0231 0.0138 0.1008 0.1008 0.1502 p, Fisher’s exact test vs empty (two-sided) uc.160+mut12 >0.9999 0.8553 0.1210 0.1210 0.3098 0.6834 0.5499 >0.9999 >0.9999 vs empty D pri-miR-376c empty vector ... pri-miR-376c (+48) pri-miR-376a-2 (+55) uc.160+ ** 100 * 100 ... % of clones WT uc.160+mut12 Edited 50 50 ... (+4) (+48) pri-miR-376a-2 0 0 EV + 12 EV + 12 60 60 empty vector ut ut .1 .1 +m +m uc uc ... 60 60 .1 .1 uc.160+ uc uc ... uc.160+mut12 ... (+11) (+15) (+55) 656 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License M. Soler et al. T-UCR uc.160+ regulates miR-376 in glioma A B U-87 MG U-87 MG RYBP (24-30) 3’UTR scr -3 W d 5’... ...3’ d miR-376a iR 6c ite ite WT m -37 WT 76 T m -37 ed miR-376a ed miR-376a 3’ 5’ edited miR-376c iR a iR 6a c WT m 376 miR-376c edited - RYBP (597-604) 3’UTR 5’... ...3’ iR ns r sc 2.5 m Relative expression miR-376a 3’ 5’ 2.0 RYBP ns RYBP (2086-2092) 3’UTR 5’... ...3’ 1.5 LAMIN B1 miR-376a 3’ 5’ 1.0 FOXP2 0.5 RYBP (446-452) 3’UTR 5’... ...3’ LAMIN B1 0.0 miR-376c 3’ 5’ RYBP FOXP2 RYBP (6491-6497) 3’UTR 5’... ...3’ miR-376c 3’ 5’ KS-1 KS-1 C -3 W ed d scr ite iR 6c it miR-376a iR 6a T 76 T RYBP (301-308) 3’UTR WT m -37 ed ed 5’... ...3’ m -37 W miR-376a edited iR a c miR-376c m 376 edited miR-376c 3’ 5’ WT miR-376c edited - iR r sc ns m RYBP (652-659) 3’UTR 5’... ...3’ 2.0 Relative expression RYBP edited miR-376c 3’ 5’ ns 1.5 LAMIN B1 FOXP2 (921-928) 3’UTR 5’... ...3’ 1.0 edited miR-376a 3’ 5’ FOXP2 0.5 FOXP2 (3663-3667) 3’UTR 5’... ...3’ LAMIN B1 0.0 edited miR-376a 3’ 5’ RYBP FOXP2 D U-87 MG E KS-1 U-87 MG KS-1 12 12 ut ut EV m EV m uc.160+ uc.160+ .1 + + 2.0 2.0 .1 + Relative expression Relative expression uc 60 + 60 uc 60 60 uc.160+ mut12 uc.160+ mut12 .1 .1 EV uc EV ns ns uc 1.5 ns ns 1.5 RYBP RYBP 1.0 1.0 FOXP2 FOXP2 0.5 0.5 LAMIN B1 LAMIN B1 0.0 0.0 RYBP FOXP2 RYBP FOXP2 F G KS-1 KS-1 U-87 MG + + U-87 MG /c c + a scr + EV + a/ 60 76 scr + uc.160+ 60 76 m + scr + EV uc iR + .1 -3 go 60 m 0 .1 -3 uc iR scr + uc.160+ go 6 antagomiR-376a/c+ uc.160+ ta .1 ta c.1 EV 2.0 EV an uc Relative expression antagomiR-376a/c+ uc.160+ an u ns Relative expression r+ r+ r+ r+ 2.0 ns sc sc sc sc 1.5 ns 1.5 ns RYBP RYBP 1.0 1.0 FOXP2 FOXP2 0.5 0.5 0.0 LAMIN B1 0.0 LAMIN B1 RYBP FOXP2 RYBP FOXP2 mature miRNAs, has not been investigated in detail. +55 on pri-miR-376a-2 are both on the corresponding One study has highlighted the weak effect of A-to-I ‘seed’ regions of miR-376c-3p and miR-376a-3p (here- editing on miR-376a-3p in terms of changing target after referred to as miR-376c and miR-376a), and so specificity [47], and we next investigated this in the can confer altered selectivity of target repertoire. To glioma cell lines. Positions +48 on pri-miR-376c and explore the impact of uc.160+-induced hyperediting of Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of 657 Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T-UCR uc.160+ regulates miR-376 in glioma M. Soler et al. Fig. 3. Impact of editing the miR-376 cluster on target regulation. (A) Predicted base-pairing between the 3’UTR of RYBP (above) and FOXP2 (below) mRNAs and the seed sequences of miR-376a-3p and miR-376c-3p (wild-type and edited). The complementarity between the seed regions and target mRNAs is boxed. (B, C) Validation of RYBP and FOXP2 as miR-376 targets was achieved in U-87 MG and KS-1 cell lines following transient transfection of synthetic mimics of hsa-miR-376a-3p or hsa-miR-376c-3p wild-type (WT), and their edited counterparts. Relative expression of endogenous mRNA levels was assessed by RT-qPCR (left), and RYBP and FOXP2 proteins were analyzed by western blot (right). Graphs present the mean  SD of three independent replicates. A Kruskal–Wallis test was used (ns = not significant). (D, E) Changes in RYBP and FOXP2 levels upon uc.160+ or uc.160+ mut12 overexpression in U-87 MG and KS-1 cell lines. Relative expression of endogenous mRNA levels was assessed by RT-qPCR (left), and RYBP and FOXP2 proteins were analyzed by western blot (right). Graphs represent the mean  SD of three independent experiments. One-way ANOVA was used (ns = not significant). (F, G) Changes in RYBP and FOXP2 levels upon uc.160+ overexpression in the presence of antagomiRs against miR-376a and c in U-87 MG and KS-1 cell lines. Relative expression of endogenous mRNA levels was assessed by RT-qPCR (left), and RYBP and FOXP2 proteins were analyzed by western blot (right). Graphs represent the mean  SD of three independent replicates. One-way ANOVA was used (ns = not significant). miR-376 on its ability to regulate downstream mRNA abolished when uc.160+ was cotransfected with antag- targets, we set out to identify bona fide miR-376a and omiRs against miR-376a and c (Fig. 3F–G, Fig. S3C), miR-376c targets in glioma cell lines. Using the Tar- and was also suppressed in KS-1 cells that had been getScan and microT-DS mRNA target prediction engineered by means of CRISPR/Cas9 to knockout tools, we found two common candidate genes for miR- the miR-376 cluster (Fig. S3D–E). Altogether, these 376a among the top 20 hits from each database: the data suggest a translational control by the ultracon- RING1 and YY1-binding protein RYBP, and the served transcript that is likely to proceed via regula- single-stranded nucleic acid-binding protein RBMS1. tion of miR-376 processing and editing. According to the miRDB prediction database, RYBP RYBP is canonically known as an epigenetic factor is also a potential target of unedited and edited forms with ubiquitin binding activity that associates with of miR-376c. In addition, the member of the forkhead/ Polycomb complexes, is required throughout develop- winged-helix family of transcription factors FOXP2 is ment, and has important roles in apoptosis and cancer one of the top newly predicted targets for edited (but [55]. Remarkably, high levels of RYBP protein induce not unedited) miR-376a (Fig. 3A). These findings apoptosis only in tumor cells, a feature of interest for prompted us to experimentally validate the regulation cancer therapy [56-59]. In fact, a lower level of RYBP of RYBP and FOXP2 by the miR-376 family. Duplex has been observed in a number of tumor types, includ- miRNA mimics were designed against the unedited ing glioblastoma, than in nontumoral tissue [60]. On and edited forms of miR-376a and miR-376c and were the other hand, FOXP2 encodes a transcription factor transfected in U-87 MG or KS-1 cells (Fig. 3B–C). with critical roles in neural development and brain cir- Neither RYBP nor FOXP2 mRNA levels were signifi- cuits controlling language acquisition. Although the cantly altered when the mimics were overexpressed; by link between FOXP2 and oncogenic features is still contrast, the encoded proteins were downregulated uncertain, it belongs to a genomic domain containing under particular conditions. Specifically, as predicted, a cluster of genes (including the MET oncogene) that RYBP was targeted in the two cell lines by wild-type often experiences cancer-associated epigenetic changes miR-376a and miR-376c, and by the edited miR-376c [61]. An analysis of FOXP2’s transcriptional targets (Fig. 3B–C). FOXP2 protein was downregulated by suggests that it may regulate the expression of pro- the unedited and the edited forms of miR-376a, oncogenic and tumor suppressor genes [61]. In the although the level of downregulation in KS-1 cells was brain, FOXP2 displays proneurogenic activities by greater with the edited miRNA, as expected (Fig. 3B– enhancing differentiation of neural precursors and C). Once confirmed that RYBP and FOXP2 are bona reducing proliferation [62], and its increased expression fide target genes regulated by the miR-376 family has been associated with a poorer clinical outcome in whose degree of specificity differs depending on the neuroblastoma [63]. Given the suggested involvement of levels of mature miRNA editing, we investigated the RYBP and FOXP2 in tumorigenic processes, the pro- impact of uc.160+ on the identified miR-376 targets. posed role of the miR-376 family as an important bio- Overexpression of uc.160+ in U-87 MG and KS-1 cells marker in gliomas [31], and our findings indicating the had a clearly negative effect on RYBP and FOXP2 epigenetic silencing of uc.160+ in glioma cell lines protein levels while not altering their mRNA levels, (Fig. 1C) and its hypermethylation in 70% of gliomas whereas cells transfected with uc.160+ mut12 remained from the TCGA cohort (Fig. 1B), we hypothesized that unaltered (Fig. 3D–E). Furthermore, this effect was the aberrant DNA methylation of uc.160+ could 658 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License M. Soler et al. T-UCR uc.160+ regulates miR-376 in glioma A B All-grade glioma 1.0 Methylated 1.0 Uc.160+ methylation (n = 438) levels in tumor 0.8 0.8 Overall Survival 0.6 0.6 0.4 0.4 Unmethylated (n = 199) 0.2 0.2 Log-rank: p < 0.001*** 0.0 HR (95% CI) = 0.122 (0.087 - 0.171) 0.0 M G LG B .0 .5 1.0 1.5 2.0 2.5 3.0 G C Years since diagnosis Glioblastoma Lower-grade glioma 1.0 Methylated 1.0 (n = 409) 0.8 0.8 Overall Survival Overall Survival 0.6 0.6 Unmethylated Unmethylated Methylated (n = 105) (n = 94) 0.4 (n = 29) 0.4 0.2 0.2 Log-rank: p = 0.234 Log-rank: p < 0.001*** HR (95% CI) = 0.708 (0.399 - 1.256) HR (95% CI) = 0.135 (0.085 - 0.213) 0.0 0.0 .0 .5 1.0 1.5 2.0 2.5 3.0 .0 .5 1.0 1.5 2.0 2.5 3.0 Years since diagnosis Years since diagnosis D Lower-grade glioma Lower-grade glioma IDH1wt IDH1mut Unmethylated 1.0 (n = 11) 1.0 Methylated 0.8 (n = 31) 0.8 Methylated Overall Survival Overall Survival (n = 378) 0.6 0.6 0.4 Unmethylated 0.4 (n = 83) 0.2 0.2 Log-rank: p = 0.002** Log-rank: p = 0.315 HR (95% CI) = 0.306 (0.136 - 0.687) HR (95% CI) Undefined 0.0 0.0 .0 .5 1.0 1.5 2.0 2.5 3.0 .0 .5 1.0 1.5 2.0 2.5 3.0 Years since diagnosis Years since diagnosis E F LGG patient characteristics (n = 503) HR (95% CI); p Age 2.96 (1.81 - 4.84); p < 0.001*** ( 60 y vs. < 60 y) Gender 1.066 (0.674 - 1.688); p = 0.784 (Male vs. Female) IDH1 mutation status 0.249 (0.125 - 0.496); p < 0.001*** (Mutated vs. WT) uc160+ methylation status 0.486 (0.248 - 0.953); p = 0.036* (Methylated vs. unmethylated) 0 1 2 3 4 5 HR for OS Benefit Risk Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of 659 Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T-UCR uc.160+ regulates miR-376 in glioma M. Soler et al. Fig. 4. uc.160+-CpG island methylation in human primary gliomas and its association with clinical outcome. (A) Levels of uc.160+ CpG island methylation in the TCGA datasets of primary lower-grade gliomas (LGG, n = 503) and high-grade gliomas (glioblastoma multiforme, GBM, n = 134). An unpaired t-test was used (****P < 0.0001). (B) Kaplan–Meier analysis of overall survival (OS) across all glioma grades from TCGA datasets with respect to uc.160+ methylation status. (C) Kaplan–Meier analysis of OS in lower- and high-grade glioma from TCGA datasets with respect to uc.160+ methylation status. (D) Kaplan–Meier analysis of OS of lower-grade glioma according to the molecular status of IDH1 gene and uc.160+ CpG island methylation levels. For all graphs in (B–D), the probabilities correspond to log-rank tests. Results of univariate Cox regression are represented as the hazard ratio (HR) and 95% confidence interval (CI). (E) Forest plot of the multivariable Cox regression of clinical outcome in the TCGA lower-grade glioma cohort on uc.160+ methylation status. Probabilities (P) and 95% of confidence intervals (95% CI) correspond to the hazard ratios (HR) associated with OS. Significant covariates were considered independent prognostic factors (*P < 0.05; ***P < 0.001). (F) Summary of results. Left, CpG island hypermethylation-associated epigenetic silencing of uc.160+ compromises the efficient biogenesis of miR-376 family members and correlates with improved overall survival in lower-grade glioma. Right, in hypomethylated samples, the expression of uc.160+ increases processing and editing of miR-376, modulating target regulation and correlating with poorer clinical outcome in LGG. Figure created with BioRender.com. influence glioma patients’ outcome via the altered regu- analyzed together (log-rank test: P < 0.001; hazard lation of miR-376 cluster and downstream target genes. ratio (HR) = 0.122, 95% CI = 0.087–0.171) (Fig. 4B); however, considering the enrichment of LGG cases (n = 503) over GBM cases (n = 134), the weight of the 3.5. Methylation of the uc.160+-associated CpG 409 methylated LGG cases could bias the analysis. island is an independent prognosis factor in Independent analysis of LGG and GBM patients lower-grade glioma demonstrated that even though methylation is not able Gliomas are the most common and lethal type of to stratify patients with dismal prognosis as glioblas- intracranial tumors and have a very poor outcome and tomas (log-rank: P = 0.234; HR = 0.708, 95% CI = a median survival of 14–16 months (for grade IV glio- 0.399–1.256), uc.160+ CpG island methylation was sig- mas). Since T-UCR expression is not available for nificantly associated with increased OS in the lower- TCGA cohorts, we investigated changes in uc.160+ pro- grade gliomas (log-rank: P < 0.001; HR = 0.135, 95% moter CpG island methylation, which we took to be a CI = 0.085–0.213) (Fig. 4C). We then examined proxy of its expression in human primary gliomas. We whether uc.160+ methylation helped define patient sur- analyzed the collections of lower-grade gliomas (LGG) vival when combined with a genetic alteration of well- and glioblastomas (GBM) from the TCGA (https:// recognized clinical impact in gliomas. Mutations in isoc- portal.gdc.cancer.gov/). DNA methylation data were itrate dehydrogenase 1 (IDH1) are present in a high available for 503 LGGs, including 242 diffuse low- percentage of lower-grade gliomas and, to a lesser grade and 261 intermediate-grade gliomas (grades II extent, in high-grade gliomas, and are markers of and III, respectively, according to the World Health improved prognosis [64]. In glioblastoma samples, Organization (WHO) classification), and 134 GBMs uc.160+ methylation did not contribute to defining the (WHO grade IV) (Table S2). DNA hypermethylation of survival of wild-type IDH1 patients (log-rank: the uc.160+ CpG island was detected in 68.75% of glio- P = 0.991; HR = 1.003, 95% CI = 0.563–1.789) (Fig. mas, with a clear enrichment in LGG (81.3% of methy- S4A). By contrast, in the case of lower-grade gliomas, lated cases) in comparison with GBM (21.6% of although methylation of uc.160+ did not contribute to methylated cases; Fisher’s exact test, P < 0.0001) (Fig. 4 patient stratification when IDH1 was mutated (in which A). The decrease detected in uc.160+ methylation as the case the prognosis is generally good), wild-type IDH1 disease progresses resembles other tumoral contexts: in patients had better OS when uc.160+ was hypermethy- colorectal cancer, a dynamic change in uc.160+ methy- lated (log-rank: P = 0.002; HR = 0.306, 95% CI =- lation has also been observed, and methylation in stage 0.136–0.687) (Fig. 4D). Finally, multivariate Cox III and IV patients has been associated with improved regression analysis including IDH1 mutational status, overall survival (OS) [21]. In addition, low methylation age, and gender showed that uc.160+ methylation is an of uc.160+ in GBM may be associated with the pro- independent prognosis factor for lower-grade glioma apoptotic role of RYBP in tumor cells [56-59] and the patients (HR = 0.486; 95% CI = 0.248–0.953; worse predicted outcome when the level of RYBP is P = 0.036) (Fig. 4E, Fig. S4B–C). reduced [60]. Considering this, we next examined To summarize, our findings demonstrate that whether uc.160+ methylation had any prognostic value uc.160+ is an epigenetically controlled T-UCR with in glioma patients. Uc.160+ methylation was associated clinical relevance in gliomas. Mechanistically, our with increased OS when gliomas of all grades were working model suggests that cells with high levels of 660 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License M. Soler et al. T-UCR uc.160+ regulates miR-376 in glioma uc.160+ expression can process and edit pri-miR-376 Conflict of interest cluster more efficiently, with an effect on key protein targets such as RYBP and FOXP2 and, importantly, The authors declare no conflict of interest. on lower-grade glioma patients’ prognosis (Fig. 4F). Overall, our data highlight the impact of the regula- Author contributions tory roles of ultraconserved transcripts as fine-tuners of other ncRNA biogenesis and their potential as All experiments were conceived by MS and SG and biomarkers in the clinical practice. Approaches based mainly carried out by MS. AS-C optimized T-UCR on HITS-CLIP have mapped globally the interaction detection and analyzed its localization. CM-M carried between miRNAs and lncRNAs and have pointed to out the search for homology between T-UCRs and miR- widespread cross-regulatory mechanisms [65,66], with NAs. VD and MCdeM performed the analysis of the an important impact on cellular physiology [67]. Given clinical parameters. FS and ES designed and prepared the complex, layered regulation that has been revealed genome-edited cells. SG wrote the manuscript with the for some T-UCRs and miRNAs [68], we expect a input of all authors. SG and ME supervised the project. bright future for these noncoding RNAs in transla- tional settings. The discovery of additional features for pri-miRNA recognition and processing and the process Peer review by which other ncRNAs influence their mature levels The peer review history for this article is available at might lead to improved diagnostic and therapeutic https://publons.com/publon/10.1002/1878-0261.13121. tools in cancer and other diseases in which miRNAs are dysregulated. Data accessibility 4. Conclusions The methylation data that support the findings of this study are available in Table S2. The conclusions derived from this work can be sum- marized as follows: References  The transcribed ultraconserved region uc.160+ 1 Yao R-W, Wang Y & Chen L-L (2019) Cellular displays a region of homology with miR-376 and functions of long noncoding RNAs. Nat Cell Biol 21, enhances its production in vitro and in cell lines. 542–551.  This is accompanied by an increase in A-to-I 2 Derrien T, Johnson R, Bussotti G, Tanzer A, Djebali S, editing on the mature miR-376, and an impact Tilgner H, Guernec G, Martin D, Merkel A, Knowles on the downstream targets RYBP and FOXP2, DG et al. (2012) The GENCODE v7 catalog of human which have roles in oncogenesis. long noncoding RNAs: analysis of their gene structure,  Methylation of uc.160+-associated CpG island in evolution, and expression. Genome Res 22, 1775–1789. glioma patients helps define survival and is an 3 Bejerano G, Pheasant M, Makunin I, Stephen S, Kent independent factor for better prognosis in lower- WJ, Mattick JS & Haussler D (2004) Ultraconserved grade glioma. elements in the human genome. Science 304, 1321–1325. 4 Calin GA, Liu C-G, Ferracin M, Hyslop T, Spizzo R, Sevignani C, Fabbri M, Cimmino A, Lee EJ, Wojcik Acknowledgements SE et al. (2007) Ultraconserved regions encoding ncRNAs are altered in human leukemias and We thank the CERCA program (Generalitat de Catalu- carcinomas. Cancer Cell 12, 215–229. nya) and the Josep Carreras Foundation for institu- 5 Braconi C, Valeri N, Kogure T, Gasparini P, Huang N, tional support. This work was supported by the Nuovo GJ, Terracciano L, Croce CM & Patel T (2011) Ministerio de Ciencia e Innovación (MCI)—Agencia Expression and functional role of a transcribed Estatal de Investigación (AEI) and the Instituto de noncoding RNA with an ultraconserved element in Salud Carlos III (ISCIII), cofinanced by the European hepatocellular carcinoma. Proc Natl Acad Sci USA 108, Development Regional Fund, ‘A way to achieve Eur- 786–791. ope’ ERDF, under grant numbers PID2019-111658RB- 6 Carotenuto P, Fassan M, Pandolfo R, Lampis A, I00 (SG), RTI2018-094049-B-I00 ‘EPIRNA’ (ME), and Vicentini C, Cascione L, Paulus-Hock V, Boulter L, the Health and Science Departments of the Catalan Guest R, Quagliata L et al. (2017) Wnt signalling Government (Generalitat de Catalunya). ME is an modulates transcribed-ultraconserved regions in ICREA Research Professor. hepatobiliary cancers. Gut 66, 1268–1277. Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of 661 Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T-UCR uc.160+ regulates miR-376 in glioma M. Soler et al. 7 Hudson Robert S, Yi M, Volfovsky N, Prueitt RL, 19 Goto K, Ishikawa S, Honma R, Tanimoto K, Esposito D, Volinia S, Liu C-G, Schetter AJ, Van Sakamoto N, Sentani K, Oue N, Teishima J, Roosbroeck K, Stephens RM et al. (2013) Transcription Matsubara A & Yasui W (2016) The transcribed- signatures encoded by ultraconserved genomic regions ultraconserved regions in prostate and gastric cancer: in human prostate cancer. Mol Cancer 12, 13. DNA hypermethylation and microRNA-associated 8 Luo H-L, Chen J, Luo T, Wu F-X, Liu J-J, Wang H-F, regulation. Oncogene 35, 3598–3606. Chen M, Li L-Q & Li H (2017) Downregulation of 20 Lujambio A, Portela A, Liz J, Melo SA, Rossi S, macrophage-derived T-UCR uc.306 associates with Spizzo R, Croce CM, Calin GA & Esteller M (2010) poor prognosis in hepatocellular carcinoma. Cell CpG island hypermethylation-associated silencing Physiol Biochem 42, 1526–1539 of non-coding RNAs transcribed from 9 Mestdagh P, Fredlund E, Pattyn F, Rihani A, Van ultraconserved regions in human cancer. Oncogene Maerken T, Vermeulen J, Kumps C, Menten B, De 29, 6390–6401. Preter K, Schramm A et al. (2010) An integrative 21 Kottorou AE, Antonacopoulou AG, Dimitrakopoulos genomics screen uncovers ncRNA T-UCR functions in F-ID, Georgia D, Chaido S, Melpomeni K, Theodoros neuroblastoma tumours. Oncogene 29, 3583–3592. T, Chrysa O, Katsakoulis EC, Angelos K et al. (2018) 10 Zhou J, Wang C, Gong W, Wu Y, Xue H, Jiang Z & Deregulation of methylation of transcribed-ultra Shi M (2018) uc.454 inhibited growth by targeting heat conserved regions in colorectal cancer and their value shock protein family A member 12B in non-small-cell for detection of adenomas and adenocarcinomas. lung cancer. Mol Ther Nucleic Acids 12, 174–183. Oncotarget 9, 21411–21428. 11 Bo C, Li N, Li X, Liang X & An Y (2016) Long 22 Michlewski G &Cáceres JF (2019) Post-transcriptional noncoding RNA uc.338 promotes cell proliferation control of miRNA biogenesis. RNA 25, 1–16. through association with BMI1 in hepatocellular 23 Lim LP, Lau NC, Garrett-Engele P, Grimson A, carcinoma. Hum Cell 29, 141–147. Schelter JM, Castle J, Bartel DP, Linsley PS & Johnson 12 Zhang Y, Wang S, Qian W, Ji D, Wang Q, Zhang Z, JM (2005) Microarray analysis shows that some Wang S, Ji B, Fu Z & Sun Y (2018) uc.338 targets p21 microRNAs downregulate large numbers of target and cyclin D1 via PI3K/AKT pathway activation to mRNAs. Nature 433, 769–773. promote cell proliferation in colorectal cancer. Oncol Rep. 24 Lewis BP, Burge CB & Bartel DP (2005) Conserved 13 Wang C, Wang Z, Zhou J, Liu S, Wu C, Huang C & seed pairing, often flanked by adenosines, indicates that Ding Y (2017) TUC.338 promotes invasion and thousands of human genes are microRNA targets. Cell metastasis in colorectal cancer. Int J Cancer 140, 1457– 120, 15–20. 1464. 25 Taccioli C, Fabbri E, Visone R, Volinia S, Calin GA, 14 Scaruffi P, Stigliani S, Moretti S, Coco S, De Vecchi C, Fong LY, Gambari R, Bottoni A, Acunzo M, Hagan J Valdora F, Garaventa A, Bonassi S & Tonini GP et al. (2009) UCbase & miRfunc: a database of (2009) Transcribed-ultra conserved region expression is ultraconserved sequences and microRNA function. associated with outcome in high-risk neuroblastoma. Nucleic Acids Res 37: D41–D48. BMC Cancer 9, 441. 26 Griffiths-Jones S (2004) The microRNA registry. 15 Vannini I, Wise PM, Challagundla KB, Plousiou M, Nucleic Acids Res 32, 109D–111. Raffini M, Bandini E, Fanini F, Paliaga G, Crawford 27 Oliveira-Mateos C, Sánchez-Castillo A, Soler M, M, Ferracin M et al. (2017) Transcribed ultraconserved Obiols-Guardia A, Piñeyro D, Boque-Sastre R, Calleja- region 339 promotes carcinogenesis by modulating Cervantes ME, Castro de Moura M, Martı́nez-Cardús tumor suppressor microRNAs. Nat Commun 8. A, Rubio T et al. (2019) The transcribed pseudogene 16 Liz J, Portela A, Soler M, Gómez A, Ling H, RPSAP52 enhances the oncofetal HMGA2-IGF2BP2- Michlewski G, Calin GA, Guil S & Esteller M (2014) RAS axis through LIN28B-dependent and independent Regulation of pri-miRNA processing by a long let-7 inhibition. Nat Commun 10. noncoding RNA transcribed from an ultraconserved 28 Guil S & Cáceres JF (2007) The multifunctional RNA- region. Mol Cell 55, 138–147. binding protein hnRNP A1 is required for processing of 17 Guo J, Fang W, Sun L, Lu Y, Dou L, Huang X, Tang miR-18a. Nat Struct Mol Biol 14, 591–596. W, Yu L & Li J (2018) Ultraconserved element uc.372 29 Kawahara Y (2012) Quantification of adenosine-to- drives hepatic lipid accumulation by suppressing miR- inosine editing of microRNAs using a conventional 195/miR4668 maturation. Nat Commun 9. method. Nat Protoc 7, 1426–1437. 18 Xiao L, Wu J, Wang J-Y, Chung HK, Kalakonda S, 30 Ludwig N, Leidinger P, Becker K, Backes C, Fehlmann Rao JN, Gorospe M & Wang J-Y (2018) Long T, Pallasch C, Rheinheimer S, Meder B, Stähler C, noncoding RNA uc.173 promotes renewal of the Meese E et al. (2016) Distribution of miRNA intestinal mucosa by inducing degradation of expression across human tissues. Nucleic Acids Res 44, microRNA 195. Gastroenterology 154, 599–611 3865–3877. 662 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License M. Soler et al. T-UCR uc.160+ regulates miR-376 in glioma 31 Huang Q, Wang C, Hou Z, Wang G, Lv J, Wang H, 45 Alon S, Mor E, Vigneault F, Church GM, Locatelli F, Yang J, Zhang Z & Zhang H (2017) Serum microRNA- Galeano F, Gallo A, Shomron N & Eisenberg E (2012) 376 family as diagnostic and prognostic markers in Systematic identification of edited microRNAs in the human gliomas. Cancer Biomark 19, 137–144. human brain. Genome Res 22, 1533–1540. 32 Auyeung VC, Igor U, McGeary SE & Bartel DP (2013) 46 Wang Y, Xu X, Yu S, Jeong KJ, Zhou Z, Han L, Beyond secondary structure: primary-sequence Tsang YH, Li J, Chen H, Mangala LS et al. (2017) determinants license Pri-miRNA hairpins for Systematic characterization of A-to-I RNA editing processing. Cell 152, 844–858. hotspots in microRNAs across human cancers. Genome 33 Kim K, Nguyen TD, Li S & Nguyen TA (2018) SRSF3 Res 27, 1112–1125. recruits DROSHA to the basal junction of primary 47 Kume H, Hino K, Galipon J & Ui-Tei K (2014) A-to-I microRNAs. RNA 24, 892–898. editing in the miRNA seed region regulates target 34 Han K, Wang F-W, Cao C-H, Ling H, Chen J-W, mRNA selection and silencing efficiency. Nucleic Acids Chen R-X, Feng Z-H, Luo J, Jin X-H, Duan J-L et al. Res 42, 10050–10060. (2020) CircLONP2 enhances colorectal carcinoma 48 Paul D, Sinha AN, Ray A, Lal M, Nayak S, Sharma invasion and metastasis through modulating the A, Mehani B, Mukherjee D, Laddha SV, Suri A et al. maturation and exosomal dissemination of microRNA- (2017) A-to-I editing in human miRNAs is enriched in 17. Mol Cancer 19. seed sequence, influenced by sequence contexts and 35 Krol J, Krol I, Alvarez CPP, Fiscella M, Hierlemann significantly hypoedited in glioblastoma multiforme. Sci A, Roska B & Filipowicz W (2015) A network Rep 7, 2466. comprising short and long noncoding RNAs and RNA 49 Kawahara Y, Megraw M, Kreider E, Iizasa H, Valente helicase controls mouse retina architecture. Nat L, Hatzigeorgiou AG & Nishikura K (2008) Frequency Commun 6, 7305. and fate of microRNA editing in human brain. Nucleic 36 Kawahara Y, Zinshteyn B, Sethupathy P, Iizasa H, Acids Res 36, 5270–5280. Hatzigeorgiou AG & Nishikura K (2007) Redirection 50 Iizasa H, Wulff B-E, Alla NR, Maragkakis M, Megraw of silencing targets by adenosine-to-inosine editing of M, Hatzigeorgiou A, Iwakiri D, Takada K, Wiedmer miRNAs. Science 315, 1137–1140. A, Showe L et al. (2010) Editing of Epstein-Barr virus- 37 Wulff B-E, Sakurai M & Nishikura K (2011) encoded BART6 microRNAs controls their dicer Elucidating the inosinome: global approaches to targeting and consequently affects viral latency. J Biol adenosine-to-inosine RNA editing. Nat Rev Genet 12, Chem 285, 33358–33370. 81–85. 51 Yang W, Chendrimada TP, Wang Q, Higuchi M, 38 Behm M & Öhman M (2016) RNA editing: a Seeburg PH, Shiekhattar R & Nishikura K (2006) contributor to neuronal dynamics in the mammalian Modulation of microRNA processing and expression brain. Trends Genet 32, 165–175. through RNA editing by ADAR deaminases. Nat 39 Nishikura K (2010) Functions and regulation of RNA Struct Mol Biol 13, 13–21. editing by ADAR deaminases. Annu Rev Biochem 79, 52 Quinones-Valdez G, Tran SS, Jun H, Bahn JH, Yang 321–349. E-W, Zhan L, Brümmer A, Wei X, Van Nostrand EL, 40 Jain M, Jantsch MF & Licht K (2019) The Editor’s I Pratt GA et al. (2019) Regulation of RNA editing by on disease development. Trends Genet 35, 903–913. RNA-binding proteins in human cells. Commun Biol 2, 41 Han L, Diao L, Yu S, Xu X, Li J, Zhang R, Yang Y, 19. Werner HMJ, Eterovic AK, Yuan Y et al. (2015) The 53 Tariq A, Garncarz W, Handl C, Balik A, Pusch O & genomic landscape and clinical relevance of A-to-I RNA Jantsch MF (2013) RNA-interacting proteins act as editing in human cancers. Cancer Cell 28, 515–528. site-specific repressors of ADAR2-mediated RNA 42 Picardi E, Manzari C, Mastropasqua F, Aiello I, editing and fluctuate upon neuronal stimulation. D’Erchia AM & Pesole G (2015) Profiling RNA editing Nucleic Acids Res 41, 2581–2593. in human tissues: towards the inosinome Atlas. Sci Rep 54 Choudhury Y, Tay FC, Lam DH, Sandanaraj E, Tang 5, 14941. C, Ang B-T & Wang S (2012) Attenuated adenosine-to- 43 Ramaswami G, Zhang R, Piskol R, Keegan LP, Deng inosine editing of microRNA-376a* promotes P, O’Connell MA & Li JB (2013) Identifying RNA invasiveness of glioblastoma cells. J Clin Invest 122, editing sites using RNA sequencing data alone. Nat 4059–4076. Methods 10, 128–132. 55 da Silva S, Carolina J, Rocı́o S & Busturia A (2018) 44 Gong J, Wu Y, Zhang X, Liao Y, Sibanda VL, Liu W Epigenetic and non-epigenetic functions of the RYBP & Guo A-Y (2014) Comprehensive analysis of human protein in development and disease. Mech Ageing Dev small RNA sequencing data provides insights into 174, 111–120. expression profiles and miRNA editing. RNA Biol 11, 56 Danen-van Oorschot AAAM, Voskamp P, Seelen 1375–1385. MCMJ, van Miltenburg MHAM, Bolk MW, Tait SW, Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of 663 Federation of European Biochemical Societies 18780261, 2022, 3, Downloaded from https://febs.onlinelibrary.wiley.com/doi/10.1002/1878-0261.13121 by Consorci de Serveis Universitaris de Catalunya, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T-UCR uc.160+ regulates miR-376 in glioma M. Soler et al. Boesen-de Cock JGR, Rohn JL, Borst J & Noteborn 66 Chi SW, Zang JB, Mele A & Darnell RB (2009). MHM (2004) Human death effector domain-associated Argonaute HITS-CLIP decodes microRNA–mRNA factor interacts with the viral apoptosis agonist Apoptin interaction maps. Nature 460, 479–486. and exerts tumor-preferential cell killing. Cell Death 67 Shan K, Jiang Q, Wang X-Q, Wang Y-N-Z, Yang H, Differ 11, 564–573. Yao M-D, Liu C, Li X-M, Yao J, Liu B et al. (2016) 57 Wen MA, Xuan Z, Meng LI, Xiaoli MA, Bingren H, Role of long non-coding RNA-RNCR3 in atherosclerosis- Hong C & Chen D (2016) Proapoptotic RYBP interacts related vascular dysfunction. Cell Death Dis 7, e2248. with FANK1 and induces tumor cell apoptosis through 68 Terreri S, Durso M, Colonna V, Romanelli A, the AP-1 signaling pathway. Cell Signal 28, 779–787. Terracciano D, Ferro M, Perdonà S, Castaldo L, 58 Novak RL & Phillips AC (2008) Adenoviral-mediated Febbraio F, de Nigris F et al. (2016) New cross-talk Rybp expression promotes tumor cell-specific apoptosis. layer between ultraconserved non-coding RNAs, Cancer Gene Ther 15, 713–722. microRNAs and polycomb protein YY1 in bladder 59 Stanton SE, Blanck JK, Joseph L & Schreiber-Agus N cancer. Genes 7, 127. (2007) Rybp interacts with Hippi and enhances Hippi- 69 Francesco I, Knijnenburg TA, Vis DJ, Bignell GR, mediated apoptosis. Apoptosis 12, 2197–2206. Menden MP, Michael S, Nanne A, Emanuel G, Syd B, 60 Li G, Warden C, Zou Z, Neman J, Krueger JS, Jain A, Howard L et al. (2016) A landscape of pharmacogenomic Jandial R & Chen M (2013) Altered expression of interactions in cancer. Cell 166, 740–754. polycomb group genes in glioblastoma multiforme. PLoS One 8, e80970. 61 Herrero MJ & Gitton Y (2018) The untold stories of Supporting information the speech gene, the FOXP2 cancer gene. Genes & Additional supporting information may be found Cancer 9, 11–38. online in the Supporting Information section at the end 62 Tsui D, Vessey JP, Tomita H, Kaplan DR & Miller FD of the article. (2013) FoxP2 regulates neurogenesis during embryonic Fig. S1. Methylation status of uc.160+ in normal tis- cortical development. J Neurosci 33, 244–258. sues, expression patterns and coding petential. 63 Khan FH, Pandian V, Ramraj S, Natarajan M, Fig. S2. Expression of miR-376 cluster membres in tis- Aravindan S, Herman TS & Aravindan N (2015) sues and cell lines, and in vitro modulation of their Acquired genetic alterations in tumor cells dictate the processing by uc.160+. development of high-risk neuroblastoma and clinical Fig. S3. ADAR1 and 2 levels in glioma cell lines, and outcomes. BMC Cancer 15, 514. disruption of miR-376 function. 64 Cohen AL, Holmen SL & Colman H (2013) IDH1 and IDH2 mutations in gliomas. Curr Neurol Neurosci Rep Fig. S4. uc.160+ CpG island hypermethylation in high- 13, 345. grade gliomas and its association with clinical out- 65 Franco-Zorrilla JM, Valli A, Todesco M, Mateos I, come. Puga MI, Rubio-Somoza I, Leyva A, Weigel D, Garcı́a Table S1. Oligos used in this work. JA & Paz-Ares J (2007) Target mimicry provides a new Table S2. DNA methylation data of 637 cases of mechanism for regulation of microRNA activity. Nat glioma (503 LGG and 134 GBMs) available in TCGA Genet 39, 1033–1037. (https://portal.gdc.cancer.gov/). 664 Molecular Oncology 16 (2022) 648–664 ª 2021 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies