Long non‐coding RNA TINCR suppresses metastatic melanoma dissemination by preventing ATF4 translation

Article15 February 2021Open Access Source DataTransparent process Long non-coding RNA TINCR suppresses metastatic melanoma dissemination by preventing ATF4 translation Marine Melixetian orcid.org/0000-0002-8454-8262 Department of Experimental Oncology, IEO, European Institute Oncology IRCCS, Milan, Italy Search for more papers this author Daniela Bossi Marija Mihailovich Simona Punzi Iros Barozzi Surgery and Cancer, Imperial College London, UK Federica Marocchi Alessandro Cuomo Tiziana Bonaldi orcid.org/0000-0003-3556-1265 Giuseppe Testa orcid.org/0000-0002-9104-0918 Hemato-oncology, University Jean-Christophe Laboratory Molecular Cancer Biology, KULeuven, Leuven, Belgium Center VIB, Eleonora Leucci Saverio Minucci Biosciences, Pier Pelicci Corresponding Author [email protected] orcid.org/0000-0002-5076-2316 Luisa Lanfrancone orcid.org/0000-0002-4523-3815 Information Melixetian1, Bossi1,8, Mihailovich1, Punzi1,9, Barozzi2, Marocchi1, Cuomo1, Bonaldi1, Testa1,3, Marine4,5, Leucci6, Minucci1,7, *,1,3 *,1 1Department 2Department 3Department 4Laboratory 5Center 6Laboratory 7Department 8Present address: Research (IOR), Bellinzona, Switzerland 9Present Candiolo Institute—FPO Torino, *Corresponding author. Tel: +39 02 57489868; E-mail: 94375011; EMBO Reports (2021)22:e50852https://doi.org/10.15252/embr.202050852 PDFDownload PDF article text main figures. Peer ReviewDownload a summary the editorial decision including letters, reviewer comments responses to feedback. ToolsAdd favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract Transition from proliferative-to-invasive phenotypes promotes metastasis therapy resistance in melanoma. Reversion invasive phenotype, however, is challenged poor understanding mechanisms underlying its maintenance. Here, we report that lncRNA down-regulated silencing increases expression levels markers, vitro migration, vivo tumor growth, BRAF MEK inhibitors. The critical mediator ATF4, central player integrated stress response (ISR), which activated TINCR-depleted cells absence starvation eIF2α phosphorylation. depletion global protein synthesis induces translational reprogramming, leading increased mRNAs encoding other ISR proteins. Strikingly, re-expression reduces numbers tumor-initiating formation, drug sensitivity. Mechanistically, interacts with associated their binding ribosomes. Thus, suppressor functions nutrient-rich conditions repressing selected RNAs. SYNOPSIS blocks acquisition melanoma, dissemination. targets RNAs phenotypes, interaction High prevent reprogramming translation, maintaining proliferative, epithelial-like states. downregulation leads activation, promoting invasion conditions. RNAs, translating Introduction Cutaneous most aggressive form skin cancer, mainly due early metastasization rapid development available treatments (Kunz Holzel, 2017; Schadendorf et al, 2018). Metastatic connected high propensity migrate invade neighboring tissue, property acquired since earliest steps melanomagenesis (Damsky 2014). Though treatment has been revolutionized recent years, thanks use targeted therapies (such as BRAFV600E or MAP-kinase inhibitors) immune checkpoint (PD-1 CTLA-4) inhibitors (Landsberg 2012; Woods 2014; Hugo 2016; Tirosh 2016b; Jaberg-Bentele Kunz 2017), rate either less than 50%, relapse near 100% ~ 30% immunotherapy 2018; Garcia-Jimenez Goding, 2019). Emerging evidence indicates result capacity dynamically change phenotype environmental perturbations, so called plasticity (Rambow Melanoma can fact reprogram transcription metabolism nutrient deprivation extracellular signals microenvironment (TME), resulting into generation distinct cell degree intratumor heterogeneity (Falletta Ferguson Gene analyses lines patient biopsies led identification two predominant transcriptional states, are linked phenotypically populations, named, respectively, “proliferative” “invasive” (Hoek 2008). proliferative related highly proliferating melanocytes characterized MITF, master regulator melanocyte proliferation differentiation, upstream SOX10 (Bondurand 2000; Lee Potterf Verastegui Goding Arnheiter, instead, relates an undifferentiated state under control gene-regulatory network orchestrated AP1 TEAD transcriptions factors (Riesenberg 2015; Verfaillie 2015). These undifferentiated/invasive low MITF tyrosine-protein kinase receptor AXL ligands like TGFβ WNT5A (Konieczkowski Muller Arozarena Wellbrock, Of note, do not correlate presence known driver mutations, NRASQ61L, indicating largely independent genetic make-up individual (Ferguson Rambow phenotypic states per se plastic convert each others changes TME, revealed existence single patterns consistent transient (Tirosh Kim A large body experimental interconversion between play major role during (“phenotype-switch” model). In particular, may drive transition generate able metastasize. Upon seeding at new sites, must revert allow growth (invasive-to-proliferative switch) (Carreira 2006; Hoek 2008; Pinner 2009; 2010; Phenotypic also allows adaptation emergence resistant disease (Arozarena 2019; Treatment (MAPKi) progressive dedifferentiated intrinsically drug-resistant possess features (expression NGFR, PDGFR, IGF1R, EGFR, AXL, ROR2) (O'Connell 2013; On hand, melanomas express similar signature (innate anti-PD1 signature; IPRES) includes (Hugo Jenkins switch, treatments, be induced experimentally using some TME involved progression, such hypoxia, (glucose glutamine), inflammation, TGFβ, WNT5A, TNF-α Widmer Falletta Vivas-Garcia 2020). Instead, reversion manipulation intracellular signaling pathway proved particularly difficult successful approaches still zebrafish model system, it recently demonstrated that, after colonization distant re-acquire differentiated secretion endothelins (Kim 2017). data suggest sites provides required induce invasive-to-proliferative targeting might represent effective therapeutic option. seems have central, though controversial modulation switches (Goding However, suggesting additional, yet poorly characterized, maintenance phenotype. Over past decade, long (lncRNAs) identified players gene regulation (Chen They defined 5′-capped 3′-polyadenylated transcripts > 200 bp, no minimal protein-coding potential. Aberrant lncRNAs human tumorigenesis, (Huarte, 2016a). epithelial tumors, several (ATB21, HOTAIR, CHRF, Hh52) regulate (ZEB, SNAIL, TWIST) epithelial-to-mesenchymal (EMT), resembling observed (Gugnoni Ciarrocchi, reported promote survival (SAMMSON) (Leucci 2016b), suppression (Coe 2019), (SPRY4-IT1, UCA1, MALAT-1, BANCR, PVT1, CASC15) (Lessard Bhan whether contribute establishment aberrant unknown. We here acts Results Transcriptional profiles distinguish primary groups nevi metastases To identify transcriptional-targets dissemination, performed unsupervised consensus clustering 89 356 Genome Atlas (TCGA) three stable clusters (Figs 1A EV1A), none NRAS mutations (Fig EV1B). Primary were distributed Clusters 1 2 (~ 35 56%, respectively; P < 0.01), while 3 71 22%, 0.01) 1A). Unsupervised clearly (Primary-A Primary-B, Fig 1B), corresponded to, Cluster 2/3 samples (31/34 Primary-A 1; 54/55 Primary-B 3) EV1C). Differential 1,034 genes (Primary Signature; fold ≥ 1.5, FDR ≤ 0.05) (Dataset EV1 EV1D). Down-regulated prevailed over up-regulated power discriminating (P = 5.9e-15 5.3e-05, respectively) 1C). same larger dataset 297 (Staaf EV2A) small primary/metastatic our tissue bank (12 samples, EV2B). Notably, latter, showed similarity melanomas, respectively EV2C D; Spearman correlation coefficient ρ 0.7). Figure 1. TCGA cutaneous RNA-seq identifies Heatmap showing (n 445: melanomas). indicated below heatmap: cluster 55, red), 304, blue), 86, green). Type (primary metastatic) site (subcutaneous, lymph node distant) within above heatmap, respectively. 89) 1,000 variable expressed dataset. divides groups: 34, red) pink), indicated. Boxplots distributions normalized rank-sums (left panel) (right signature. better discriminate 5.9e-15, Wilcoxon rank-sum test). Median, minimum, maximum, 25th, 75th whiskers distribution shown. Bar-graph Set Enrichment Analysis (GSEA) differentially signature), combining GO sets (GO) set enrichment analysis Download figure PowerPoint Click expand figure. EV1. Consensus matrix produced k-means (k matrix: matrix. Frequencies mutations. Upper panel: statistically significant association mutation frequencies was found (χ2 test value 0.01). Lower 2) 89). square: 34) 55). Volcano plot Genes false discovery (FDR) 0.05 |Fold Change| 1.5 shown red. EV2. GEO IEO datasets GSE57715 297) based on previously see Figs 1B EV1C) melanomas. hierarchical cohort (IEO dataset) 3), 4) 5) Bar plots (left) (right) component compared processes potentially relevant positive regulators migration integrin-mediated cell-cell adhesion, among EV2, 1D). enriched epidermal differentiation (cell-cell apical junction assembly), WNT canonical pathway, SOX10-driven program Together, these indicate grouped categories, transcriptome similarities differences alternative influencing (Primary-A; nevi-like) invasive/migratory programs (Primary-B; metastasis-like), correlates non-metastatic then searched implicated inhibition switch e.g., metastases. annotated signature, namely TINCR, KIAA0125 MIAT. Only significantly (Mann–Whitney test, 0.001) 2A). correlated (Spearman’s 0.79, 3.6e-3) 2B). Down-regulation neither variations copy number EV3A), nor hypermethylation CpG dinucleotides flanking start (TSS, EV3B). QPCR obtained 8 group) grown immunodeficient mice (patient-derived xenografts; PDXs; n 4, 8, metastasis) lower all 2C). matched pairs PDXs markedly reduced counterparts whereas non-invasive 2. Box (from 5th 95th percentile) depicting SKCM samples. higher reduction **P values Scatter (upper (lower strongly rank value: genes. Dot qRT–PCR (8) surgical specimens patient-derived xenografts (4 PDXs). relative L32P housekeeping Paired marked asterisk. difference EV3. assays Boxplot level different aberrations (see Materials Methods details). No transcript (Kruskal–Wallis 0.1). DNA methylation CpGs (TSS) well percentile lines. (mean ± s.d.) biological replicates Representative images transwell assay. WM902B MMC70 migrated through chamber stained crystal violet five representative microscopic fields counted. Relative respective “negative” (shSCR transduced cells). Scale bars: 100 μm. Collagen assay spheroid cultures. shTINCR infected 24 96 h culture collagen I. Mean (±s.e.m.) 9 11 spheroids group condition. Statistical Student t-test (**P 0.01, ***P 0.001). invasive-phenotype next investigated effects cells, focusing migration/proliferation, tumorigenicity, sensitivity, transcriptomic patterns. used (WM902B MMC70) relatively 3A EV3C), expressing (as proteins EPHA2, 3G). Infection lentiviral vectors carrying TINCR-specific (shTINCR-1 shTINCR-2) shRNAs (> 75%), (shSCR) 3A). 3. Loss inducing quantification lines, independently shTINCR-2). Cells shScamble control. measured experiments both knockdown efficiency 70% shRNAs. shTINCR-1, shTINCR-2, shSCR. i