TGF-β orchestrates fibrogenic and developmental EMTs via the RAS effector RREB1

Abstract

Epithelial-to-mesenchymal transitions (EMTs) are phenotypic plasticity processes that confer migratory and invasive properties to epithelial cells during development, wound-healing, fibrosis and cancer^1-4. EMTs are driven by SNAIL, ZEB and TWIST transcription factors^5,6 together with microRNAs that balance this regulatory network^7,8. Transforming growth factor β (TGF-β) is a potent inducer of developmental and fibrogenic EMTs^4,9,10. Aberrant TGF-β signalling and EMT are implicated in the pathogenesis of renal fibrosis, alcoholic liver disease, non-alcoholic steatohepatitis, pulmonary fibrosis and cancer^4,11. TGF-β depends on RAS and mitogen-activated protein kinase (MAPK) pathway inputs for the induction of EMTs^12-19. Here we show how these signals coordinately trigger EMTs and integrate them with broader pathophysiological processes. We identify RAS-responsive element binding protein 1 (RREB1), a RAS transcriptional effector^20,21, as a key partner of TGF-β-activated SMAD transcription factors in EMT. MAPK-activated RREB1 recruits TGF-β-activated SMAD factors to SNAIL. Context-dependent chromatin accessibility dictates the ability of RREB1 and SMAD to activate additional genes that determine the nature of the resulting EMT. In carcinoma cells, TGF-β-SMAD and RREB1 directly drive expression of SNAIL and fibrogenic factors stimulating myofibroblasts, promoting intratumoral fibrosis and supporting tumour growth. In mouse epiblast progenitors, Nodal-SMAD and RREB1 combine to induce expression of SNAIL and mesendoderm-differentiation genes that drive gastrulation. Thus, RREB1 provides a molecular link between RAS and TGF-β pathways for coordinated induction of developmental and fibrogenic EMTs. These insights increase our understanding of the regulation of epithelial plasticity and its pathophysiological consequences in development, fibrosis and cancer.

Extended Data Figure 1.. RREB1 as a SMAD cofactor in TGF-β gene responses

(a) YFP fluorescence images of CIY organoids expressing KRAS^G12D under doxycycline control treated with SB or TGF-β for 2.5 days. Scale bars, 200 μm. Images are representative of two independent experiments, (b) Influence of KRAS^G12D on TGF-β gene responses. CIY pancreatic organoids inducibly expressing KRAS^G12D, were treated with SB or TGF-β for 1.5 h and subjected to RNA-seq analysis. Dots represent Log2 fold change in mRNA levels of individual genes under TGF-β versus SB treatment conditions, with KRAS^G12D expression turned off (X axis) or on (Y axis). Off-diagonal dots correspond to TGF-β gene responses that were enabled (groups I and II) or disabled (groups III and IV) by KRAS^G12D. Gene activation (I and III) and repression responses (II and IV) are included, (c) Heatmap representation of four classes of KRAS-modified TGF-β gene responses. n=1. Representative result of two independent experiments. Classes I-IV correspond to the off-diagonal genes derived from the RNA-seq in (b). (d) TGF-β gene activation responses augmented by KRAS^G12D (class I responses) in CIY pancreatic organoids. Fold-increase in mRNA levels in TGF-β vs. SB treatment conditions, in presence or absence of inducible KRAS^G12D. (e) Heatmaps showing TGF-β induction of Snai1, Has2, Il11, Smad7 and Skil in four independent CIY mouse pancreatic organoid lines with inducible KRAS^G12D expression. n=4. (f) Heatmap representation of the indicated TGF-β gene responses in spheroid cultures of pancreatic epithelial cells (PECs) inducibly expressing KRAS^G12D. n=2. (g) Heatmap representation of the indicated TGF-β gene responses in monolayer cultures of mouse Kras^G12D;Smad4^fl/fl;Cdkn2a^fl/fl;Pdx1-Cre (KSIC) PDA cell lines transduced with a SMAD4 vector or an empty vector. n=2. (h) Transcription factor (TF)-binding motifs enriched in KRAS-independent SMAD2/3 binding sites (left panel) and KRAS-dependent SMAD2/3 binding sites (right panel). SMAD2/3 ChIP-seq analyses were performed in SMAD4-restored PDA cells that were treated with SB (2.5 μM) or TGF-β (100 pM) for 1.5 h. Transcription factor binding motif analyses were performed with PscanChIP. n=821 peak regions (left panel). n=778 peak regions (right panel), (i) Motif enrichment analysis of RAS-regulated transcription factors in KRAS-dependent (n=778 peak regions) and KRAS-independent (n=821 peak regions) SMAD2/3 binding sites. (j) Comparative enrichment of classic SMAD binding motifs (CAGAC, GGCTG) and 5GC motifs (GGC(GC)|(CG)) in 200 bp-region of SMAD2/3 ChIP peaks within 1000 bp from a transcriptional start site. The relative enrichment is normalized to the baseline data set obtained from 20,000 random 200 bp regions from the mm10 genome assembly. The 5GC motifs are ~4-fold enriched in SMAD2/3 ChIP peaks compared to the baseline, and the classic motifs are 2-fold enriched.

Extended Data Figure 2.. RREB1 interacts with SMAD and binds to TGF-β target genes

(a) Western immunoblot analysis of RREB1 and HA-RREB1 levels in SMAD4-restored PDA cells stably transduced with an HA-RREB1 vector. Anti-Tubulin immunoblotting was used as loading control. Data are representative of two independent experiments, (b) Proximity ligation assay showing TGF-β dependent proximity between RREB1, SMAD2/3 and SMAD4 in the nucleus. Scale bars, 30 μm. Data are representative of two independent experiments, (c) Quantification of PLA signals in (b). Cell numbers (n) of each group is indicated in the graph, two-tailed unpaired t test. Center values and error bars: mean ± s.d. ****, p<0.0001; ***, p<0.001. (d) SMAD4-restored PDA cells expressing HA-RREB1 were treated with TGF-β for 1.5 h, lysed, and immunoprecipitated (IP) with the indicated antibodies. The immune complexes were collected and subjected to western immunoblot with the antibodies indicated on the left. Data are representative of two independent experiments. (e) Heatmap representation of ChIP-seq tag densities for SMAD2/3 and HA-RREB1 in genomic regions ±3 kb from the center of SMAD2/3 binding peaks in SMAD4-restored PDA cells that were treated with SB or TGF-β for 1.5 h and subjected to SMAD2/3 and HA-RREB1 ChIP-seq analysis. ChIP-seq was performed once and an independent ChIP was performed in which selective genomic regions were confirmed by qPCR.

Extended Data Figure 3.. RREB1 is phosphorylated and regulated by ERK

(a) Representative immunofluorescence images of HA-RREB1 in SMAD4-restored PDA cells treated with DMSO or 1 μM ERK inhibitor SCH772984 (ERKi) for 6 h. Scale bar, 20 μm. Data are representative of two independent experiments, (b) (c) Western immunoblot analysis of RREB1 (b) or HA-RREB1 levels (c) in SMAD4-restored PDA cells treated with DMSO, or 1 μM ERKi or 1 μM AZD6244 (MEKi), which is an inhibitor of the ERK-activating kinases MEK1/2 for the indicated time periods. Anti-Tubulin immunoblotting was used as loading control. Data are representative of two independent experiments, (d) ChIP-PCR analysis of HA-RREB1 binding to the indicated sites (refer to Fig. 1 g, 3b) in Snai1, Has2, and Il11 in SMAD4-restored PDA cells that were treated with vehicle DMSO or the ERKi (1 μM) for 6 h. Mean ± s.e.m. n=3, two-way ANOVA analysis. **, p<0.01; ***, p<0.001; ****, p<0.0001. (e) SMAD4-restored PDA cells expressing HA-RREB1 were treated with ERKi for the indicated length of time. HA-REBB1 was tested for binding to Snai1 DE2 and Has2 PP1 dsDNA oligonucleotide probes in DNA affinity precipitation assays. Data are representative of two independent experiments, (f) Schematic representation of RREB1. Each tick represents a previously annotated phosphorylation site in PhosphoSitePlus® identified in at least two independent mass-spectrometry experiments. Red filled circles represent high stoichiometry (>15%) phosphorylation sites that are inhibited by ERKi, as identified in (g). Zinc-finger domains annotated in Uniprot are shown. (g) Phosphorylation stoichiometry of four ERK-dependent RREB1 phosphorylation sites in SMAD4-restored PDA cells, as determined by SILAC mass spectrometry of cells treated with DMSO (control) in light medium or ERKi in heavy medium for 6 h. (h) Summary of ERK-dependent RREB1 phosphorylation sites, sequence motifs, and phosphorylation stoichiometry, (i) RREB1 KO PDA cells were transduced with the indicated RREB1 WT or phosphorylation site mutant constructs, then treated with SB or TGF-β for 1.5 h. mRNA levels of Snai1 and Has2 were determined by qRT-PCR. Mean ± s.e.m. n=4, two-way ANOVA analysis. ***, p<0.001; ****, p<0.0001. (j) ChIP-PCR analysis of HA-RREB1 binding to the indicated sites in RREB1 KO PDA cells transduced with the indicated RREB1 WT or phosphorylation site mutant constructs. Mean ± s.e.m. n=4, two-tailed unpaired t test. ****, p<0.0001.

Extended Data Figure 4.. RREB1 mediates KRAS-dependent TGF-β responses in PDA cells

(a) Scheme of CRISPR/Cas9-mediated mutation of Rreb1 in mouse SMAD4-restored PDA cells. (b) sgRNA sequences and genomic sequences of Rreb1 coding region (CDS) exons 1 and 7 in mutant clones K01 and K02 derived from SMAD4-restored PDA cells, (c) Western immunoblot analysis of RREB1 levels in Rreb1 wild type (WT) and knockout (KO) cells. Anti-Tubulin immunoblotting was used as loading control. Data are representative of two independent experiments, (d) Western immunoblot analysis of E-cadherin in mouse KSIC PDA cells, SMAD4-restored PDA cells, and two RREB1 KO SMAD4-restored PDA clones, treated with SB or TGF-β for 24 h. Anti-Tubulin immunoblotting was used as loading control. Data are representative of two independent experiments, (e) Representative E-cadherin and DAPI immunofluorescence images of the same cells as in (f) treated with SB or TGF-β for 48 h. Scale bars, 100 pm. Data are representative of two independent experiments, (f) Gene track view of SMAD2/3 ChIP-seq tags in the Smad7 locus of the RREB1 WT and KO PDA cells. The gene body is schematically represented at the bottom. ChIP-seq was performed once and an independent ChIP was performed in which selective genomic regions were confirmed by qPCR. (g) mRNA levels of Snai1, Has2 and Il11 in WT and two RREB1 KO cells that were transduced with an RREB1 vector or empty vector and then treated with SB or TGF-β for 1.5 h. Mean ± s.e.m. n=4, two-way ANOVA analysis, ****, p<0.0001.

Extended Data Figure 5.. RREB1 mediates tumorigenic EMT in lung adenocarcinoma cells

(a) Snai1, Has2 and Il11 mRNA levels in 393T3 mouse LUAD cells treated with DMSO (Ctrl) or ERKi (SCH772984, 1 μM) for 6 h, followed with treatment of SB or TGF-β for 1.5 h. Mean ± s.e.m. n=4, two-way ANOVA analysis, ****, p<0.0001. (b) Cdh1 mRNA levels in 393T3 cells with the indicated treatments for 48 h. Mean ± s.d. n=4, two-way ANOVA analysis, (c) Western immunoblot analysis of E-cadherin in 393T3 cells with the indicated treatments for 48 h. Anti-Tubulin immunoblotting was used as loading control. Data are representative of two independent experiments, (d) SMAD4-restored PDA cells and 393T3 LUAD cells cultured in D10F containing 2.5 μM MK2206 were treated with SB (2.5 μM) or TGF-β (100 pM) and assayed for cleaved caspase 3/7 activity at the indicated times. Mean ± s.e.m. n=4, two-way ANOVA analysis, ***, p<0.001. (e) SMAD4-restored PDA cells and 393T3 cells cultured in D10F containing 2.5 μM MK2206 were treated with SB or TGF-β. Cell viability was determined at the indicated times. Mean ± s.e.m. n=4, two-way ANOVA analysis, ***, p<0.001. (f) sgRNA sequence targeting Rreb1 CDS exon3, and mutant Rreb1 genomic sequences of the resulting 393T3 KO1 and KO2 clones, (g) mRNA levels of Snai1 and Has2 in the RREB1 WT and KO 393T3 cells after treatment with SB (2.5 μM) or TGF-β (100 pM) for 1.5 h. Mean ± s.e.m. n=4, two-way ANOVA analysis. ****, p<0.0001. (h) Phase contrast images of 393T3 cell monolayers treated with SB or TGF-β for 48 h. Scale bars, 200 μm. Data are representative of two independent experiments, (i) Weight and volume of tumours in Fig. 2g. Mean ± s.e.m. n=10, two sites were inoculated per mouse, two-tailed unpaired t test. ****, p<0.0001. (j) Representative hematoxylin and eosin staining images of indicated lung tissue sections in Fig. 2h. Scale bars, 200 μm. Data are representative of two independent experiments.

Extended Data Figure 6.. RREB1-dependent TGF-β responses in LUAD and PDA cells

(a) sgRNA sequence targeting RREB1 CDS exon3, and mutant RREB1 genomic sequences of the resulting A549 KOI and KO2 clones, (b) SNAIL and SLUG mRNA levels in WT A549 and two RREB1 KO clones treated with SB or TGF-β for 24 h. Mean ± s.e.m. n=4, two-way ANOVA analysis. ****, p<0.0001. (c) Phase contrast images of WT A549 and RREB1 KO cell monolayers treated with SB or TGF-β for 48 h. Scale bars, 200 μm. Data are representative of two independent experiments. (d) Growth kinetics of tumours formed by subcutaneously inoculated WT or RREB1 KO A549 cells in athymic mice, as determined by BLI of a transduced firefly luciferase gene in the cells. Mean ± s.e.m. n=10, two sites were inoculated per mouse, two-way ANOVA analysis, (e) Gene ontology analysis of TGF-β response genes in CIY organoids inducibly expressing KRAS^G12D, based on the RNA-seq in Extended Data Fig. 1b. (f) RREB1 WT and KO PDA cells were treated with SB or TGF-β for 1.5 h and subjected to RNA-seq analysis. Dots represent Log2 (fold change) in mRNA levels of individual genes under TGF-β versus SB treatment conditions, in RREB1 KO (X axis) or WT cells (Y axis). Off-diagonal dots corresponding to Snai1, Has2, Il11 and Wisp1 are highlighted, (g) Induction of Snai1 and Zeb1 expression by TGF-β in mouse PDA cells. Mean ± s.d. n=4. (h) sgRNA sequence targeting Snai1 and resulting mutant Snai1 genomic sequences in mouse PDA cells (i) Knockdown of Zeb1 with two independent shRNAs in SNAIL KO mouse PDA cells (KOsh cells). Mean ± s.d. n=4. (j) Fibrogentic gene responses to TGF-β in WT and SNAIL/ZEB1-doubly depleted KOsh PDA cells. Mean ± s.d. n=4. (k-m) E-cadherin levels (k), phase contrast images (I), and E-cadherin and Zeb1 immunofluorescence staining in WT and KOsh PDA cells that were treated with SB or TGF-β for 48h. Scale bars, 100 μm. Data are representative of two independent experiments.

Extended Data Figure 7.. RREB1 dependent TGF-β responses in mammary epithelial cells

(a) sgRNA sequence targeting RREB1 CDS exon3, and mutant RREB1 genomic sequences of the resulting NMuMG KO1 and KO2 clones, (b) Phase contrast images of RREB1 WT and KO NMuMG cell monolayers treated with SB or TGF-β for 48 h. Scale bar, 100 μm. Data are representative of two independent experiments, (c) Western immunoblot analysis of E-cadherin in RREB1 WT and KO NMuMG cells, treated with SB or TGF-β for 48 h. Anti-β-actin immunoblotting was used as loading control. Data are representative of two independent experiments, (d) RREB1 WT and KO NMuMG cells were treated with SB or TGF-β for 1.5 h and subjected to RNA-seq analysis. Dots represent Log2 fold change in mRNA levels of individual genes under TGF-β versus SB treatment conditions, in RREB1 KO (X axis) or WT cells (Y axis). Off-diagonal dots corresponding to Snai1 and Has2 are highlighted, (e) ChIP-PCR analysis of SMAD2/3 binding to the Snai1 (DE2) and Has2 (UE1) regions (refer to Fig. 1g) in RREB1 WT and KO NMuMG cells. Cells were treated with 2.5 μM SB or 100 pM TGF-β for 1.5 h. Mean ± s.e.m. n=4. two-way ANOVA analysis. ***, p<0.001; ****, p<0.0001. (f) mRNA levels of Snai1 and Has2 in RREB1 WT and KO NMuMG cells after treatment with SB or TGF-β for 1.5 h. Mean ± s.e.m. n=4. two-way ANOVA analysis. ****, p<0.0001. (g) ChIP-PCR analysis of HA-RREB1 binding to the indicated Snai1 and Has2 regions in NMuMG cells that were treated with vehicle DMSO or the ERKi SCH772984 (1 μM) for 6 h. Mean ± s.e.m. n=3, two-tailed unpaired t test, (h) Snai1 and Has2 mRNA levels in NMuMG cells treated with DMSO (Ctrl), ERKi (1 μM SCH772984), EGF (10 ng/ml, 10 min), or EGFR inhibitor (Gefitinib, 1 μM, 2 h), followed by SB or TGF-β treatment for another 1.5 h. Mean ± s.e.m. n=4. two-way ANOVA analysis. ***, p<0.001; ****, p<0.0001.

Extended Data Figure 8.. RREB1 in gastrulation EMT and mesendoderm differentiation

(a) Corn plot presentation of Rreb1, Snai1, Cdh2, Gsc and Brachyury/T in E7.0 mouse embryo. A: anterior, L: left, R: right, P: posterior regions. Each dot represents transcript level at a specific positional address. Heatmap denotes expression level of each gene computed from transcript counts in RNA-seq datasets, (b) Reads per million reads (RPM) of Rreb1, Snai1, Twist1, Cdh2, Eomes and Zeb2 in the RNA-seq dataset at the indicated times after shifting ESCs into LIF-deficient EB differentiation media, (c) sgRNA sequence targeting Rreb1 CDS exon3, and mutant Rreb1 genomic sequences of four resulting mESC KO clones, (d) mRNA levels of EMT (Snai1, Cdh2) and mesendoderm differentiation genes (Eomes, Gsc, T/Brachyury and Mixi1) in WT and four independent RREB1 KO clones on Day4 EB differentiation. Mean ± s.d., n=4, two-way ANOVA analysis. ****, p<0.0001. (e) mRNA levels of the indicated genes in WT and four independent RREB1 KO clones treated with receptor inhibitor (SB) or Activin A (AC) for 2 h. Mean ± s.e.m. n=4, two-way ANOVA analysis. ****, p<0.0001. (f) GSEA for gastrulation, EMT and stem cell differentiation genes in WT cells, and absence in RREB1 KO cells, at Day4 EB differentiation.

Extended Data Figure 9.. RREB1 and SMAD contextually regulate EMT genes

(a) Heatmap representation of ChIP-seq tag densities for SMAD2/3 and HA-RREB1 in genomic regions ±3 kb from center of 3422 high-confidence SMAD2/3 binding peaks in Day3 EBs subjected to SMAD2/3 and HA ChIP-seq analyses, (b) Gene track view of SMAD2/3 and HA-RREB1 ChIP-seq tags in the loci of EMT genes (Has2, Twist1, and ZebT) and early mesendoderm lineage genes (Eomes, Brachyury/T, and Mixi1) in Day-3 EBs. Gene bodies are schematically represented at the bottom of each track set. (c) Gene track view of ATAC-seq, and SMAD2/3 and RREB1 ChIP-seq tags on indicated loci, in Day3 EBs (red tracks) versus TGF-β treated (1.5 h) SMAD4-restored PDA cells (blue tracks), (a-c) ChIP-seq was performed once and an independent ChIP was performed in which selective genomic regions were confirmed by qPCR.

Extended Data Figure 10.. Rreb1^−/− mouse embryo chimeras exhibit defects in early development

(a) E7.5 and E8.5 chimeric embryos containing WT ESCs or Rreb1^−/− ESCs were scored, based on gross morphology, as normal/mild defects, developmentally retarded or severely abnormal. At E7.5, a fraction of Rreb1^+/+ ESC embryos displayed small clumps of cells in the amniotic cavity, possibly an artifact from the microinjection, and hence were scored as abnormal. Rreb1^+/+ data is compiled from 4 distinct KO clones, (b) Images showing brightfield morphology and mCherry fluorescence (marking descendants of injected ESCs) in representative litters of Rreb1^+/+ ESC-containing chimeric embryos dissected at E7.5 and E8.5. nc, non-chimeric; lc, low chimerism. Asterisks mark morphologically abnormal/developmentally retarded embryos, (c) Brightfield images of morphologically abnormal Rreb1^−/− ESC-containing chimeric E8.5 embryos. Embryos exhibited abnormal headfold development including disproportionate headfolds (i), asymmetric headfolds (ii). Axis duplication was also observed (iii) and (iv). To note, the embryo in panel (iii) is also developmentally retarded. (d)(e) Confocal maximum intensity projections of wholemount immunostained E8.5 Rreb1^−/− ESC-containing chimeric embryos. Panel (d), an embryo with an ectopic somite-like structure (arrowhead). Panel (e), the embryo in (c)(iv) with axis duplication of the headfolds. (f) Sagittal confocal optical sections of wholemount immunostained chimeric E7.5 embryos. Embryos shown in (f)(i)-(ii) have multiple cavities and multiple expression sites of SNAIL hence anterior-posterior axis orientation is not possible, (g) Brightfield images of morphologically abnormal Rreb1^−/− ESC-containing chimeric E7.5 embryos. Embryos frequently had protrusions into the cavity and thickening of the posterior epiblast, marked by arrowheads. (h)(i) Confocal maximum intensity projections of chimeric embryos after wholemount immunostaining for phospho-Histone H3 (h), labeling mitotic cells, and cleaved Caspase 3 (i) labeling apoptotic cells. Brackets demarcate the primitive streak, (j) Sagittal confocal optical sections of chimeric E7.5 embryos after wholemount immunostaining for E-cadherin and N-cadherin. Arrowhead, aberrant N-cadherin expression. HF, headfold; PS, primitive streak; A, anterior; P, posterior; Pr, proximal; Ds, distal; L, left; R, right. Scale bars, 50 μm. (b-j) Images are representative of two independent experiments.

Figure 1.. RREB1; a KRAS-dependent SMAD cofactor

(a) Schematic of source and generation of CIY pancreatic epithelial organoids and SMAD4-restored PDA cells. (b) Snai1 and Smad7 mRNA levels in pancreatic epithelial organoid cultures. Cells engineered to express KRAS^G12D under doxycycline control treated with TGF^/Nodal receptor inhibitor SB505124 (SB, 2.5 μM) or TGF-β (10 pM) for 1.5 h. Mean ± s.d. n=4, two-way ANOVA analysis, ****, p<0.0001. (c) E-cadherin, ZEB1 and DAPI immunofluorescence images of CIY pancreatic organoids +/− KRAS^G12D treated with SB or TGF-β for 2.5 days. Scale bars, 30 μm. Images are representative of two independent experiments. (d) Screening of pancreatic progenitor transcription factor shRNA library for mediators of TGF-β-induced lethal EMT. Dot plot of shRNA enrichment in TGF-β-treated versus SB-treated SMAD4-restored PDA cells. Sox4 and Rreb1 transcription factors scoring positive in the screen. shRNAs targeting Tgfbr1 and Tgfbr2 included as positive controls. (e) Position of RREB1 peak summits relative to summits of overlapping SMAD2/3 peaks (left), and position of SMAD2/3 peak summits relative to summits of overlapping RREB1 peaks (right), based on ChIP-seq analysis in Extended Data Figure 2d. (f) Venn diagram depicting overlap between SMAD2/3 and RREB1 ChIP-seq peaks, based on ChIP-seq analysis in Extended Data Figure 2d. (g) Gene track view of SMAD2/3 and HA-RREB1 ChIP-seq tags at indicated loci and experimental conditions. Gene bodies represented at bottom of track sets. PP: proximal promoter; DE: downstream enhancer; UE: upstream enhancer. ChIP-seq was performed once and an independent ChIP was performed in which selective genomic regions were confirmed by quantitative PCR (qPCR). See also Extended Data Figure 1–3 and Supplemental Information Movie 1.

Figure 2.. RREB1 mediates KRAS and TGF-β dependent EMT

(a) Gene track view of SMAD2/3 ChIP-seq tags at indicated loci of RREB1 WT and KO SMAD4-restored mouse PDA cells. Gene bodies represented at the bottom of track sets. DE: downstream enhancer, UE: upstream enhancer. ChIP-seq was performed once and an independent ChIP was performed in which selected genomic regions were confirmed by qPCR. (b) ChIP-PCR analysis of SMAD2/3 binding to indicated sites of Snai1 (DE2) and Has2 (UE1) in RREB1 WT and KO PDA cells after treatment with SB (2.5 μM) or TGF-β (100 pM) for 1.5 h. Mean ± s.e.m. n=4, two-way ANOVA analysis, ****, p<0.0001. (c) Levels of Snai1, Has2, Il11, and Smad7 in RREB1 WT and KO PDA cells after treatment with SB (2.5 μM) or TGF-β (100 pM) for 1.5 h. Mean ± s.e.m. n=4, two-way ANOVA analysis, ****, p<0.0001. (d) Volume of RREB1 WT and KO SMAD4-restored PDA tumours after subcutaneous inoculation in syngeneic FVB mice. Mean ± s.e.m. n=10 tumours, 5 mice per group, two-way ANOVA analysis. (e,f) Representative hematoxylin and eosin staining (e), cleaved caspase-3 immunohistochemistry (IHC) (f) and SNAIL IHC (g) images of subcutaneous tumours formed by RREB1 WT and KO SMAD4-restored PDA cells 35 days after inoculation. Scale bars (e-f) upper panels, 50 μm; (e-f) lower panels, 10 μm; (g) 50 μm. (e-g) Images are representative of five biological replicates. (h) Quantification of cleaved caspase-3-positive and SNAIL-positive cells in PDA tumour sections. n=5 per group, two-tailed unpaired t test. ****, p<0.0001. Violin plots: midline, median; dotted lines, 25% and 75% quartiles. (i) Images of subcutaneous tumours formed by RREB1 WT or KO 393T3 lung adenocarcinoma cells in syngeneic B6129SF1/J mice excised 35 days after inoculation. Scale bars of left panel, 10 mm. Tumour growth monitored by firefly luciferase bioluminescence imaging (BLI) plotted over time (right panel). Mean ± s.e.m. n=10 tumours, 5 mice per group, two-way ANOVA analysis, (j) Representative ex vivo brightfield and BLI of lungs from mice inoculated via tail vein to test lung colonizing activity of RREB1 WT or KO 393T3 cells. Lungs excised and imaged 21 days after inoculation. Lung colonization load was determined by quantitative BLI. Mean ± s.e.m. n=6 mice per group, two-tailed unpaired t test. See also Extended Data Figure 4–6.

Figure 3.. RREB1 mediates a TGF-β fibrogenic response

(a) Heatmap of fibrogenic gene responses in RREB1 WT and KO PDA cells after treatment with SB or TGF-β for 1.5 h. n=2. (b) Gene track view of SMAD2/3 and HA-RREB1 ChIP-seq tags at indicated loci and experimental conditions. Gene bodies represented at bottom of track sets. PP: proximal promoter; UE: upstream enhancer. ChIP-seq was performed once and an independent ChIP was performed in which selective genomic regions were confirmed by qPCR. (c) Heatmap representation of fibrogenic genes in RREB1 WT and KO 393T3 cells treated with SB or TGF-β for 1.5 h. n=4. (d) Representative images of α-SMA and PDGFRB IHC, Masson’s trichrome stain, and α-SMA/GFP immunofluorescence of colonized lung tissue after tail vein injection of WT or Rreb1 KO GFP+ 393T3 cells. Scale bars, 100 μm. (e) Quantification of staining in (d). n for each group indicated in graph, two-tailed unpaired t test. ****, p<0.0001; ***, p<0.001. Violin plots: all data points, midline, median; dot lines, 25% and 75% quartiles. See also Extended Data Figures 6 and 7.

Figure 4.. RREB1 and SMAD regulate distinct context-dependent EMTs

(a) Heatmap representing transcripts up- or down-regulated during embryoid body (EB) differentiation. RNA-seq was performed at indicated times after shifting ESCs into differentiation medium (−LIF). EMT (red) and mesendoderm lineage genes (blue) are highlighted. n=2. (b) Gene set enrichment analysis (GSEA) for EMT, stem cell differentiation and gastrulation signatures in Day 4 EBs. (c) Transcripts exhibiting extensive up- or down-regulation (fold change > 4 or < 0.25) in WT and RREB1 KO cells, on Day4 relative to Day0 of differentiation. n=2. (d) Gene track view of ATAC-seq, and SMAD2/3 and RREB1 ChIP-seq tags at indicated loci, in Day3 EBs (red tracks) versus TGF-β treated (1.5 h) PDA cells (blue tracks). ATAC-seq and ChIP-seq were performed once. Independent ATAC and ChIP were performed in which selected genomic regions were confirmed by qPCR. (e) Chimeras generated by injecting wild-type (WT) Rreb1^+/+ or mutant Rreb1^−/− mCherry tagged ESCs into WT mouse blastocysts were transferred to pseudopregnant females and dissected at E7.5-E8.5. (f,h) Brightfield images of WT and Rreb1^−/− chimeric embryos at E8.5 (f) and E7.5 (h). Rreb1^−/− chimeras displayed morphological defects. Arrowheads in (f), somites. Arrowheads in (h), abnormal accumulation of cells within epiblast. (g,i,j) Confocal images of wholemount immunostained chimeras. (g) Maximum intensity projections of E8.5 chimeras. Rreb1^−/− chimera shows abnormal neurectoderm development and axis duplication (double allantois). (i) Sagittal section showing Brachyury expression in multiple regions and extensive epiblast folding and multiple cavities in Rreb1^−/− chimera (right panel). Due to abnormal morphology, anterior-posterior orientation of the embryo was not possible, (j) Sagittal sections of whole chimeras and representative sections through primitive streak region. Arrowheads, abnormal epiblast folding. Yellow dashed lines, boundary between epiblast and mesoderm. Brackets, primitive streak. HF, headfold; NT, neural tube; Al, allantois; Epi, epiblast; PS, primitive streak; ExM, extraembryonic mesoderm; meso, mesoderm; A, anterior; P, posterior; Pr, proximal; Ds, distal; L, left; R, right. Scale bars, 50 μm. (f-j) Images are representative of two independent experiments. (k) Summary of RAS-dependent TGF-β or Nodal effects, coordinately triggered by cooperation between RREB1 and SMAD2/3 to activate EMT and associated contextual programs in carcinoma progenitors and pluripotent embryonic cells. Principal RREB1/SMAD2/3 target genes in each program and context are indicated. See also Extended Data Figures 8–10.