Cooperative genetic networks drive embryonic stem cell transition from naïve to formative pluripotency

• DOI: 10.15252/embj.2020105776
• Publication Date: 2021-03-09T11:01:43Z
• Authors: Andreas Lackner; Robert Sehlke; Marius Garmhausen; Giuliano Giuseppe...; Michelle Huth; Fabian Titz‐Teixeira; Petra van der Lelij; Julia Ramesmayer; Henry F Thomas; Meryem Ralser; Laura Santini; Elena Galimberti; Mihail Sarov; A Francis Stewart; Austin Smith; Andreas Beyer; Martin Leeb
• Keywords: Excellence

Resource9 March 2021Open Access Transparent process Cooperative genetic networks drive embryonic stem cell transition from naïve to formative pluripotency Andreas Lackner orcid.org/0000-0003-1168-7947 Max Perutz Laboratories Vienna, University of Vienna Biocenter, AustriaThese authors contributed equally this work as first authors. Search for more papers by author Robert Sehlke Cologne Excellence Cluster Cellular Stress Response in Aging-Associated Diseases (CECAD), Cologne, GermanyThese Marius Garmhausen orcid.org/0000-0002-8617-388X Giuliano Giuseppe Stirparo orcid.org/0000-0002-5911-8682 Wellcome - MRC Cambridge Stem Cell Institute, Cambridge, UK Living Systems Exeter, UKThese Michelle Huth orcid.org/0000-0002-1152-9140 work. Fabian Titz-Teixeira orcid.org/0000-0002-7007-9039 Petra van der Lelij orcid.org/0000-0002-7461-9645 Austria Julia Ramesmayer Henry F Thomas Meryem Ralser Laura Santini orcid.org/0000-0001-9968-2459 Elena Galimberti Mihail Sarov Planck Institute Molecular Biology and Genetics, Dresden, Germany A Francis Stewart Biotechnology Center, Center Bioengineering, Technische Universität Austin Smith orcid.org/0000-0002-3029-4682 Beyer Corresponding Author [email protected] orcid.org/0000-0002-3891-2123 Medicine (CMMC), Martin Leeb orcid.org/0000-0001-5114-4782 Information Lackner1,8, Sehlke2, Garmhausen2, Stirparo3,4, Huth1, Titz-Teixeira2, Lelij1, Ramesmayer1, Thomas1, Ralser3, Santini1, Galimberti1, Sarov5, Stewart5,6, Smith3,4, *,2,7 *,1 1Max 2Cologne 3Wellcome 4Living 5Max 6Biotechnology 7Center 8Present address: Anatomy Biology, Medical *Corresponding author. Tel: [phone]; E-mail: [phone]; The EMBO Journal (2021)40:e105776https://doi.org/10.15252/embj.2020105776 PDFDownload PDF article text main figures. Peer ReviewDownload a summary the editorial decision including letters, reviewer comments responses feedback. ToolsAdd favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract In mammalian embryo, epiblast cells must exit state acquire pluripotency. This is recapitulated mouse (ESCs), which undergo progression defined conditions vitro. However, our understanding molecular cascades gene involved remains fragmentary. Here, we employed combination screens haploid ESCs, CRISPR/Cas9 disruption, large-scale transcriptomics computational systems biology delineate regulatory circuits governing exit. Transcriptome profiles 73 ESC lines deficient regulators predominantly manifest delays on trajectory epiblast. We find that operative ESCs are also active during pre- post-implantation utero. identified 496 state-associated genes tightly connected vivo largely conserved primate embryos. Integrated analysis mutant transcriptomes revealed funnelling multiple activities into discrete modules. Finally, how intersections with signalling pathways direct pivotal transition. SYNOPSIS Combining transposon-based screen biology, resource identifies controlling (ESC) pluripotency, delineates involved. Transcriptional characterization differentiation-defective defines modules inputs essential Genetic depletion specific results increased similarity pre-implantation Identification establishes an extended network transcription factor Klf2 only weakly wired network. Introduction Mouse (ESCs) can self-renew (Smith, 2017). provides amenable experimental system dissection fate paradigm (Buecker et al, 2014; Kalkan Naïve under control (GRN) containing core factors (TFs) Pou5f1, Sox2 naïve-specific TFs such Nanog, Esrrb, Klf4 others (Chen 2008; Dunn Niwa 2018). culture include inhibitors against Mek1/2 (PD0325901) Gsk3 (CHIR990201, CH; collectively termed "2i"), be homogenously maintained (Ying 2008). Within 24–36 h after withdrawal 2i, transit entirely losing identity (Kalkan During transition, GRN extinguished expression Otx2, Pou3f1, Dnmt3a/b Fgf5 initiated. similar evident peri-implantation development, where TF maintaining dissolves between day (E) 4.5 E5.5 (Boroviak Acampora 2016; Mohammed speed notable because (i) cycle around 12 long; (ii) all required establish maintain expressed robustly cells; (iii) recursively self-reinforcing. rapid dissolution implies existence circuit-breaking mechanisms. recent years, have various promoting differentiation using diploid ES (Guo 2011; Wutz, Betschinger 2013; Li Robust assays employing expressing Rex1 promoter-driven destabilised GFP reporter (Rex1::GFPd2) enable high resolution 2017; Mulas Rex1-GFP downregulation initiated within 24h 2i (N24) completed 48h (N48). Nevertheless, exact nature, mechanistic underpinnings sequence events remain partially understood. particular, lack insight different components co-operate elicit proper driven saturation, thus providing extensive list utilised information approach explore principles To evaluate dependencies causal relationships circuitries, probed response programme comprehensive series knockouts. Through integration profiling data trajectories, expose foundations choice at junction early development. Results Haploid saturation efficient platform insertional mutagenesis-based (Elling Kokubu Takeda, 2014). previously reported medium-scale comprising approximately 5 × 104 mutagenic identify regulating (Leeb now assaying 1.2 million mutations receptive genomic regions cause two independent Rex1-reporter (Fig 1A B), utilising three transposon vectors 35 1C D). Figure 1. Establishment KO informed Illustration Rex1-GFPd2 line its (in short Rex1-GFP) linked Shutdown indicates commitment differentiation. FACS levels throughout 72h time course withdrawal. Scheme screening strategy candidate After random mutagenesis piggyBac gene-trap vectors, were released Cells exposure isolated insertion sites mapped. cumulative number hits (red) novel (blue) shown. Representative plots showing 24 Tcf7l1, Rbpj, Trim71, Smg5 Pten KOs. Myc served negative control. Blue KO, dashed indicate WT. Differential N24 versus WT RC9 cells. Black dots show significance (FDR ≤ 0.05, H0: |log2FC| < log2(1.5)). Pluripotency red dots, orange blue green dots. t-SNE projection KOs based 3068 differentially strength delay observed respective indicated colour gradient measured average marker log2 fold change (log2 FC, Tfcp2l1, Tbx3, Prdm14 Klf4, Zfp42) N24. Red: delayed differentiation; blue: accelerated Similar (G) Download figure PowerPoint Stringent filtering resulted 489 (Dataset EV1). These generated inventory machinery mediates Reassuringly, known Fgfr2, Jarid2 Mapk1 (Erk2) Smith, 2014) among highest ranked (Appendix Fig S1A). hit enriched processes regulation, epigenetics signalling-related functions S1B C; Dataset EV1), well RNA-binding in-line emerging evidence RNA mechanisms (Ye Blelloch, Many not common processes, implying utilises widely cellular machinery. Therefore, mediating might other processes. library systematic transcriptional characterise deficiencies detail, selected genes, top screen. included protein complexes members recovered, even if just below cut-off threshold (e.g. Paf complex member Leo1, mTORC1 regulator Tsc2 NMD component Smg6), Mbd3, Zfp281 L3mbtl3 players (Betschinger 2013). Three either (Nestin), expected neutral (Hprt) or whose ablation was accelerate (c-Myc). Paired gRNAs used disrupt target biparental Rex1::GFPd2 carrying Cas9 transgene (henceforth cells) 1A) (Li Following parallelised approach, established passage matched isogenic homozygous lines, maximising comparability S1D E). All validated PCR, followed Sanger sequencing when required. Full deficiency 14 (Eed, Suz12, Jarid2, Kdm6a, Smg5, Smg6, Smg7, Tsc2, Pten, Raf1, Nmt1 Csnk1a1) S2A). Only heterozygote clones could (Erk2), resulting reduced levels. Notably, Erk1 Erk2 heterozygous failed rescue strong Mapk1het For five further (Alg13, Dido1, Msi2, Etv5, Jmjd1c), confirmed absence corresponding transcripts out frame deletion exon RT–qPCR RT–PCR products. Successful experiments 3xflag-tagged transgenes six (Rbpj, Fgfr1, Mbd3 Tcf7l1) causality genotype phenotype S2B C). Thus, tested knockouts showed impact expression. cannot exclude possibility hypomorphic phenotypes some cases. behaviour highly dependent density timing medium changes. robust comparison performed parallel differentiations batch wise duplicates, always controls across seven experiments. At N24, assayed status 1E) extracted transcriptome analysis. already poised Batch-corrected RNA-seq S3A B) replicates (DEGs) (H0: log2(1.5), FDR 0.05; 1F, Appendix S4A–C Interestingly, most EV6) did significantly present DEGs 1F S4C), 21% differential (6% up-, 15% downregulated). embedded ground decommissioning entry facilitate interrogation datasets, developed interactive online tool (GENEES—Genetic Network Explorer Embryonic Cells—http://shiny.cecad.uni-koeln.de:3838/stemcell_network/). Using t-Distributed Stochastic Neighbour Embedding (t-SNE), visualised similarities DEG 1G H). clustering same pathway: Eed-Suz12 (PRC2), Ptpn11-Raf1-Fgfr1-Etv5 (Fgf/ERK), Smg5-Smg6 (NMD; nonsense mediated decay), Mta3-Mbd3 (NuRD) Pten-Tsc2 (mTORC1 signalling) clustered genotype, but mainly condition (2i N24) S4A). consistent observation despite manifesting ultimately departed longer courses, loss Rex1-GFP. Furthermore, displayed globally adjusted towards knockout single sufficient permanently block culture, accord finding ternary Etv5 Rbpj sustained self-renewal LIF 2019). exceptional role Csnk1a1 involvement compensatory samples S4A), indicating special mutant. N48 downregulated, illustrating S5A). siRNA treatment Epiblastin A, chemical inhibitor (Ursu 2016), without apparently affecting proliferation duration assay S5B–D). serine threonine kinase beta-catenin destruction complex. Although another Apc, downstream repressor Tcf7l1 upregulation gene-sets S5E), stronger defects larger amplitude deregulation independently derived mutants. these mutants exhibited markedly impaired N2B27-based media S5F). Upon continuous (~5 passages), restored potential regained, suggesting likely effect rate kinetics, complicating characterisation. second case adaptation Pum1 pronounced passages S5G), seen acute CRISPR lost later passages, WT-like 2A S5G). 2. Systematic (top) (bottom) changes compared ESCs. Clustering shows wide distribution phenotypes. transfection Klf2-specific siRNAs. Plot (KO WT) mean extent global (defined correlation log2FCs N24). Each dot corresponds one Pearson's shown plot. Comparison deregulated Differentiation coded according WT, Jmjd1c cultured CH. feedback wiring do not, general, reduce (TF) state. moderate significant increases while aforementioned limited both Tbx3 EV1A). Ctbp2 upregulated Nr0b1 Other had no effects 2i. (Dunn Niwa, 2018) neutralisation (Martello contrast, markers. Several lower baseline Fgf5, Pou3f1 (Oct6) EV1B). In-line results, noted several Fgf/ERK Dnmt3a/b, effectively inhibited 2008), suggest residual pathway activity moonlighting mediate Click here expand figure. EV1. RNA-seq-derived relative (colour scale FDR, 0.05 shown). N30, transfected log2(1.5)) log2FC (phenotype strength). Correlation regarding gene. plotted. Red line: total least square regression; regression coefficients ten defective 2A). overall correlated, appeared exception. notably whereas it unaffected KOs, comparable genes. uncoupled Forced